diff --git "a/MI300x_llama8b_pp1_tp1/profiling_bs1_pl128.json" "b/MI300x_llama8b_pp1_tp1/profiling_bs1_pl128.json" new file mode 100644--- /dev/null +++ "b/MI300x_llama8b_pp1_tp1/profiling_bs1_pl128.json" @@ -0,0 +1,18162 @@ +{ + "context": { + "python_version": "3.12.10 (main, Apr 9 2025, 08:55:05) [GCC 11.4.0]", + "torch_version": "2.7.0a0+git295f2ed", + "engine_args": { + "model": "deepseek-ai/DeepSeek-R1-Distill-Llama-8B", + "served_model_name": null, + "tokenizer": null, + "hf_config_path": null, + "task": "auto", + "skip_tokenizer_init": false, + "tokenizer_mode": "auto", + "trust_remote_code": false, + "allowed_local_media_path": "", + "download_dir": null, + "load_format": "dummy", + "config_format": "auto", + "dtype": "auto", + "kv_cache_dtype": "auto", + "seed": null, + "max_model_len": null, + "cuda_graph_sizes": [ + 512 + ], + "distributed_executor_backend": null, + "pipeline_parallel_size": 1, + "tensor_parallel_size": 1, + "data_parallel_size": 1, + "enable_expert_parallel": false, + "max_parallel_loading_workers": null, + "block_size": null, + "enable_prefix_caching": null, + "prefix_caching_hash_algo": "builtin", + "disable_sliding_window": false, + "disable_cascade_attn": false, + "use_v2_block_manager": true, + "swap_space": 4, + "cpu_offload_gb": 0, + "gpu_memory_utilization": 0.9, + "max_num_batched_tokens": 8000, + "max_num_partial_prefills": 1, + "max_long_partial_prefills": 1, + "long_prefill_token_threshold": 0, + "max_num_seqs": null, + "max_logprobs": 20, + "disable_log_stats": false, + "revision": null, + "code_revision": null, + "rope_scaling": {}, + "rope_theta": null, + "hf_token": null, + "hf_overrides": {}, + "tokenizer_revision": null, + "quantization": null, + "enforce_eager": true, + "max_seq_len_to_capture": 8192, + "disable_custom_all_reduce": false, + "tokenizer_pool_size": 0, + "tokenizer_pool_type": "ray", + "tokenizer_pool_extra_config": {}, + "limit_mm_per_prompt": {}, + "mm_processor_kwargs": null, + "disable_mm_preprocessor_cache": false, + "enable_lora": null, + "enable_lora_bias": false, + "max_loras": 1, + "max_lora_rank": 16, + "fully_sharded_loras": false, + "max_cpu_loras": null, + "lora_dtype": "auto", + "lora_extra_vocab_size": 256, + "long_lora_scaling_factors": null, + "enable_prompt_adapter": null, + "max_prompt_adapters": 1, + "max_prompt_adapter_token": 0, + "device": "auto", + "num_scheduler_steps": 1, + "multi_step_stream_outputs": true, + "ray_workers_use_nsight": false, + "num_gpu_blocks_override": null, + "num_lookahead_slots": 0, + "model_loader_extra_config": {}, + "ignore_patterns": null, + "preemption_mode": null, + "scheduler_delay_factor": 0.0, + "enable_chunked_prefill": null, + "disable_chunked_mm_input": false, + "guided_decoding_backend": "xgrammar", + "guided_decoding_disable_fallback": false, + "guided_decoding_disable_any_whitespace": false, + "guided_decoding_disable_additional_properties": false, + "logits_processor_pattern": null, + "speculative_config": null, + "qlora_adapter_name_or_path": null, + "show_hidden_metrics_for_version": null, + "otlp_traces_endpoint": null, + "collect_detailed_traces": null, + "disable_async_output_proc": false, + "scheduling_policy": "fcfs", + "scheduler_cls": "vllm.core.scheduler.Scheduler", + "override_neuron_config": {}, + "override_pooler_config": null, + "compilation_config": null, + "worker_cls": "auto", + "worker_extension_cls": "", + "kv_transfer_config": null, + "kv_events_config": null, + "generation_config": "auto", + "enable_sleep_mode": false, + "override_generation_config": {}, + "model_impl": "auto", + "calculate_kv_scales": false, + "additional_config": null, + "enable_reasoning": null, + "reasoning_parser": "", + "use_tqdm_on_load": true, + "pt_load_map_location": "cpu" + }, + "prompt_len": 0, + "batch_size": 1, + "num_steps": 2, + "complete_num_requests_per_step": null, + "save_chrome_traces_folder": null + }, + "prefill": { + "metadata": { + "num_running_seqs": null + }, + "summary_stats": [ + { + "entry": { + "name": "LlamaForCausalLM", + "cuda_time_us": 13295.749000000002, + "pct_cuda_time": 96.57240655775047, + "invocations": 1 + }, + "children": [ + { + "entry": { + "name": "VocabParallelEmbedding(weight=bfloat16[128256, 4096])", + "cuda_time_us": 5.968, + "pct_cuda_time": 0.0433479996002222, + "invocations": 1 + }, + "children": [ + { + "entry": { + "name": "void at::native::(anonymous namespace)::indexSelectLargeIndex(at::cuda::detail::TensorInfo, at::cuda::detail::TensorInfo, at::cuda::detail::TensorInfo, int, int, unsigned int, unsigned int, long)", + "cuda_time_us": 5.968, + "pct_cuda_time": 0.0433479996002222, + "invocations": 1 + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cuda_time_us": 13283.612000000001, + "pct_cuda_time": 96.4842506141935, + "invocations": 32 + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cuda_time_us": 403.05800000000016, + "pct_cuda_time": 2.927573395252407, + "invocations": 64 + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::rms_norm_kernel(c10::BFloat16*, c10::BFloat16 const*, c10::BFloat16 const*, float, int, unsigned long, unsigned long)", + "cuda_time_us": 5.487, + "pct_cuda_time": 0.03985430191126327, + "invocations": 1 + }, + "children": [] + }, + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cuda_time_us": 397.57100000000014, + "pct_cuda_time": 2.887719093341144, + "invocations": 63 + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cuda_time_us": 5713.644000000001, + "pct_cuda_time": 41.5005090194055, + "invocations": 32 + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cuda_time_us": 1277.295, + "pct_cuda_time": 9.277510581328052, + "invocations": 32 + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x48x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA2048_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT4_3_MO40_NTn1_NTA0_NTB0_NTC0_NTD0_NTM0_NEPBS8_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW4_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA4_VWB1_WSGRA0_WSGRB0_WS64_WG16_4_4", + "cuda_time_us": 1277.295, + "pct_cuda_time": 9.277510581328052, + "invocations": 32 + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cuda_time_us": 665.5940000000003, + "pct_cuda_time": 4.834478626995694, + "invocations": 32 + }, + "children": [ + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cuda_time_us": 508.1140000000001, + "pct_cuda_time": 3.6906376455876853, + "invocations": 128 + }, + "children": [] + }, + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cuda_time_us": 157.48, + "pct_cuda_time": 1.143840981408008, + "invocations": 32 + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cuda_time_us": 2453.3199999999997, + "pct_cuda_time": 17.819456162737453, + "invocations": 32 + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cuda_time_us": 171.466, + "pct_cuda_time": 1.2454269603638908, + "invocations": 32 + }, + "children": [] + }, + { + "entry": { + "name": "attn_fwd", + "cuda_time_us": 2281.8540000000003, + "pct_cuda_time": 16.574029202373566, + "invocations": 32 + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cuda_time_us": 1317.4349999999995, + "pct_cuda_time": 9.569063648344288, + "invocations": 32 + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x32x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA1024_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_2_MO40_NTn1_NTA5_NTB0_NTC0_NTD7_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB1_WSGRA0_WSGRB0_WS64_WG32_4_2", + "cuda_time_us": 1317.4349999999995, + "pct_cuda_time": 9.569063648344288, + "invocations": 32 + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cuda_time_us": 7166.909999999999, + "pct_cuda_time": 52.05616819953559, + "invocations": 32 + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cuda_time_us": 4190.567, + "pct_cuda_time": 30.437784289662257, + "invocations": 32 + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT96x128x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT6_2_MO40_NTn1_NTA4_NTB0_NTC0_NTD4_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB2_WSGRA0_WSGRB0_WS64_WG16_16_1", + "cuda_time_us": 4190.567, + "pct_cuda_time": 30.437784289662257, + "invocations": 32 + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cuda_time_us": 392.1899999999999, + "pct_cuda_time": 2.848634712334306, + "invocations": 32 + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cuda_time_us": 392.1899999999999, + "pct_cuda_time": 2.848634712334306, + "invocations": 32 + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cuda_time_us": 2584.153, + "pct_cuda_time": 18.76974919753904, + "invocations": 32 + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT128x64x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB1024_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_4_MO40_NTn1_NTA4_NTB0_NTC0_NTD0_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB4_WSGRA0_WSGRB0_WS64_WG64_4_1", + "cuda_time_us": 2439.7690000000002, + "pct_cuda_time": 17.72102976485163, + "invocations": 32 + }, + "children": [] + }, + { + "entry": { + "name": "Cijk_SB_BiasS_HAS_ScaleAlphaVec_PostGSU4_VW4", + "cuda_time_us": 144.38400000000004, + "pct_cuda_time": 1.0487194326874136, + "invocations": 32 + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cuda_time_us": 6.169, + "pct_cuda_time": 0.044807943956731024, + "invocations": 1 + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cuda_time_us": 6.169, + "pct_cuda_time": 0.044807943956731024, + "invocations": 1 + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "LogitsProcessor", + "cuda_time_us": 262.761, + "pct_cuda_time": 1.9085394978140058, + "invocations": 1 + }, + "children": [ + { + "entry": { + "name": "void at::native::(anonymous namespace)::indexSelectSmallIndex(at::cuda::detail::TensorInfo, at::cuda::detail::TensorInfo, at::cuda::detail::TensorInfo, int, int, unsigned int, long)", + "cuda_time_us": 3.004, + "pct_cuda_time": 0.021819267895286105, + "invocations": 1 + }, + "children": [] + }, + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cuda_time_us": 259.757, + "pct_cuda_time": 1.8867202299187196, + "invocations": 1 + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Sampler", + "cuda_time_us": 209.138, + "pct_cuda_time": 1.5190539444355344, + "invocations": 1 + }, + "children": [ + { + "entry": { + "name": "Memcpy DtoD (Device -> Device)", + "cuda_time_us": 31.479999999999997, + "pct_cuda_time": 0.22865198180546156, + "invocations": 7 + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<512, 1, at::native::gpu_kernel_impl(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#7}::operator()() const::{lambda(float)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#7}::operator()() const::{lambda(float)#1} const&)::{lambda(int)#1})", + "cuda_time_us": 3.083, + "pct_cuda_time": 0.022393076871227387, + "invocations": 1 + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl > >(at::TensorIteratorBase&, at::native::BinaryFunctor > const&)::{lambda(int)#2}>(int, at::native::gpu_kernel_impl > >(at::TensorIteratorBase&, at::native::BinaryFunctor > const&)::{lambda(int)#2})", + "cuda_time_us": 13.54, + "pct_cuda_time": 0.09834650043347999, + "invocations": 1 + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::(anonymous namespace)::cunn_SoftMaxForward<4, float, float, float, at::native::(anonymous namespace)::SoftMaxForwardEpilogue>(float*, float const*, int)", + "cuda_time_us": 64.616, + "pct_cuda_time": 0.4693321618914138, + "invocations": 1 + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::(anonymous namespace)::cunn_SoftMaxForward<4, float, float, float, at::native::(anonymous namespace)::LogSoftMaxForwardEpilogue>(float*, float const*, int)", + "cuda_time_us": 35.932, + "pct_cuda_time": 0.2609886597914183, + "invocations": 1 + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<512, 1, at::native::gpu_kernel_impl(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#4}::operator()() const::{lambda(long)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#4}::operator()() const::{lambda(long)#1} const&)::{lambda(int)#1})", + "cuda_time_us": 5.167, + "pct_cuda_time": 0.03753001238846315, + "invocations": 1 + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::index_elementwise_kernel<128, 4, at::native::gpu_index_kernel >(at::TensorIteratorBase&, c10::ArrayRef, c10::ArrayRef)::{lambda(char*, char const*, long)#1}>(at::TensorIteratorBase&, c10::ArrayRef, c10::ArrayRef, at::native::index_kernel_impl >(at::TensorIteratorBase&, c10::ArrayRef, c10::ArrayRef)::{lambda(char*, char const*, long)#1} const&)::{lambda(int)#1}>(long, at::native::gpu_index_kernel >(at::TensorIteratorBase&, c10::ArrayRef, c10::ArrayRef)::{lambda(char*, char const*, long)#1}>(at::TensorIteratorBase&, c10::ArrayRef, c10::ArrayRef, at::native::index_kernel_impl >(at::TensorIteratorBase&, c10::ArrayRef, c10::ArrayRef)::{lambda(char*, char const*, long)#1} const&)::{lambda(int)#1})", + "cuda_time_us": 6.649, + "pct_cuda_time": 0.0482943782409312, + "invocations": 1 + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::reduce_kernel<512, 1, at::native::ReduceOp, unsigned int, long, 4, 4> >(at::native::ReduceOp, unsigned int, long, 4, 4>)", + "cuda_time_us": 44.025, + "pct_cuda_time": 0.31977139450398495, + "invocations": 1 + }, + "children": [] + }, + { + "entry": { + "name": "Memcpy DtoH (Device -> Host)", + "cuda_time_us": 4.646, + "pct_cuda_time": 0.03374577850915421, + "invocations": 1 + }, + "children": [] + } + ] + } + ], + "model_stats": [ + { + "entry": { + "name": "LlamaForCausalLM", + "cpu_time_us": 34502.858, + "cuda_time_us": 13295.749000000002, + "pct_cuda_time": 96.57240655775047, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "VocabParallelEmbedding(weight=bfloat16[128256, 4096])", + "cpu_time_us": 78.978, + "cuda_time_us": 5.968, + "pct_cuda_time": 0.0433479996002222, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void at::native::(anonymous namespace)::indexSelectLargeIndex(at::cuda::detail::TensorInfo, at::cuda::detail::TensorInfo, at::cuda::detail::TensorInfo, int, int, unsigned int, unsigned int, long)", + "cpu_time_us": 0, + "cuda_time_us": 5.968, + "pct_cuda_time": 0.0433479996002222, + "trace": "index_select(bfloat16[128256, 4096], 0, int64[128]) <- embedding(bfloat16[128256, 4096], int64[128], -1, False, False)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 1731.725, + "cuda_time_us": 393.14300000000003, + "pct_cuda_time": 2.855556737069396, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 94.992, + "cuda_time_us": 5.487, + "pct_cuda_time": 0.03985430191126327, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::rms_norm_kernel(c10::BFloat16*, c10::BFloat16 const*, c10::BFloat16 const*, float, int, unsigned long, unsigned long)", + "cpu_time_us": 0, + "cuda_time_us": 5.487, + "pct_cuda_time": 0.03985430191126327, + "trace": "_C::rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 1296.556, + "cuda_time_us": 156.856, + "pct_cuda_time": 1.1393086168385478, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 277.424, + "cuda_time_us": 41.852, + "pct_cuda_time": 0.3039880159632204, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x48x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA2048_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT4_3_MO40_NTn1_NTA0_NTB0_NTC0_NTD0_NTM0_NEPBS8_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW4_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA4_VWB1_WSGRA0_WSGRB0_WS64_WG16_4_4", + "cpu_time_us": 0, + "cuda_time_us": 41.852, + "pct_cuda_time": 0.3039880159632204, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 6144]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 6144]) <- linear(bfloat16[128, 4096], bfloat16[6144, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 109.914, + "cuda_time_us": 21.347, + "pct_cuda_time": 0.15505190138504413, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.646, + "pct_cuda_time": 0.03374577850915421, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False) <- clone(bfloat16[128, 4096], 0) <- contiguous(bfloat16[128, 4096], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.685, + "pct_cuda_time": 0.026765646535995107, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False) <- clone(bfloat16[128, 1024], 0) <- contiguous(bfloat16[128, 1024], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 5.006, + "pct_cuda_time": 0.036360604222304346, + "trace": "_C::rotary_embedding(int64[128], bfloat16[128, 4096], bfloat16[128, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.285, + "pct_cuda_time": 0.031123689391245327, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.725, + "pct_cuda_time": 0.02705618272634512, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 723.981, + "cuda_time_us": 75.231, + "pct_cuda_time": 0.5464332034055489, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 4.606, + "pct_cuda_time": 0.033455242318804194, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[128], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "attn_fwd", + "cpu_time_us": 0, + "cuda_time_us": 70.625, + "pct_cuda_time": 0.5129779610867448, + "trace": "_attention(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], int32[2], int32[2], 128, 128, True, 0.08838834764831845, None, None, None) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 117.355, + "cuda_time_us": 18.426, + "pct_cuda_time": 0.13383549608473427, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x32x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA1024_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_2_MO40_NTn1_NTA5_NTB0_NTC0_NTD7_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB1_WSGRA0_WSGRB0_WS64_WG32_4_2", + "cpu_time_us": 0, + "cuda_time_us": 18.426, + "pct_cuda_time": 0.13383549608473427, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 4096]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 4096]) <- linear(bfloat16[128, 4096], bfloat16[4096, 4096], None)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 49.644, + "cuda_time_us": 7.05, + "pct_cuda_time": 0.0512070035491901, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 7.05, + "pct_cuda_time": 0.0512070035491901, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 262.431, + "cuda_time_us": 223.75, + "pct_cuda_time": 1.625186814770395, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 99.218, + "cuda_time_us": 129.492, + "pct_cuda_time": 0.9405528090201027, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT96x128x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT6_2_MO40_NTn1_NTA4_NTB0_NTC0_NTD4_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB2_WSGRA0_WSGRB0_WS64_WG16_16_1", + "cpu_time_us": 0, + "cuda_time_us": 129.492, + "pct_cuda_time": 0.9405528090201027, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 28672]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 28672]) <- linear(bfloat16[128, 4096], bfloat16[28672, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 31.187, + "cuda_time_us": 11.777, + "pct_cuda_time": 0.08554111784380308, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 11.777, + "pct_cuda_time": 0.08554111784380308, + "trace": "_C::silu_and_mul(bfloat16[128, 14336], bfloat16[128, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 108.361, + "cuda_time_us": 82.48100000000001, + "pct_cuda_time": 0.5990928879064892, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT128x64x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB1024_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_4_MO40_NTn1_NTA4_NTB0_NTC0_NTD0_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB4_WSGRA0_WSGRB0_WS64_WG64_4_1", + "cpu_time_us": 0, + "cuda_time_us": 77.995, + "pct_cuda_time": 0.5665092541587351, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + }, + { + "entry": { + "name": "Cijk_SB_BiasS_HAS_ScaleAlphaVec_PostGSU4_VW4", + "cpu_time_us": 0, + "cuda_time_us": 4.486, + "pct_cuda_time": 0.03258363374775415, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 1083.638, + "cuda_time_us": 419.552, + "pct_cuda_time": 3.047375993343235, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 39.589, + "cuda_time_us": 6.248, + "pct_cuda_time": 0.045381752932672305, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.248, + "pct_cuda_time": 0.045381752932672305, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 758.142, + "cuda_time_us": 188.353, + "pct_cuda_time": 1.3680840765249083, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 104.346, + "cuda_time_us": 40.52, + "pct_cuda_time": 0.29431316082456493, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x48x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA2048_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT4_3_MO40_NTn1_NTA0_NTB0_NTC0_NTD0_NTM0_NEPBS8_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW4_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA4_VWB1_WSGRA0_WSGRB0_WS64_WG16_4_4", + "cpu_time_us": 0, + "cuda_time_us": 40.52, + "pct_cuda_time": 0.29431316082456493, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 6144]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 6144]) <- linear(bfloat16[128, 4096], bfloat16[6144, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 78.968, + "cuda_time_us": 21.988, + "pct_cuda_time": 0.1597077438354031, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.445, + "pct_cuda_time": 0.03228583415264539, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False) <- clone(bfloat16[128, 4096], 0) <- contiguous(bfloat16[128, 4096], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.965, + "pct_cuda_time": 0.028799399868445208, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False) <- clone(bfloat16[128, 1024], 0) <- contiguous(bfloat16[128, 1024], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 5.087, + "pct_cuda_time": 0.03694894000776312, + "trace": "_C::rotary_embedding(int64[128], bfloat16[128, 4096], bfloat16[128, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.486, + "pct_cuda_time": 0.03258363374775415, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.005, + "pct_cuda_time": 0.029089936058795226, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 439.456, + "cuda_time_us": 76.432, + "pct_cuda_time": 0.5551565525208082, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 4.646, + "pct_cuda_time": 0.03374577850915421, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[128], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "attn_fwd", + "cpu_time_us": 0, + "cuda_time_us": 71.786, + "pct_cuda_time": 0.5214107740116539, + "trace": "_attention(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], int32[2], int32[2], 128, 128, True, 0.08838834764831845, None, None, None) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 99.098, + "cuda_time_us": 49.413, + "pct_cuda_time": 0.3589066193441319, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x32x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA1024_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_2_MO40_NTn1_NTA5_NTB0_NTC0_NTD7_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB1_WSGRA0_WSGRB0_WS64_WG32_4_2", + "cpu_time_us": 0, + "cuda_time_us": 49.413, + "pct_cuda_time": 0.3589066193441319, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 4096]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 4096]) <- linear(bfloat16[128, 4096], bfloat16[4096, 4096], None)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 38.106, + "cuda_time_us": 6.369, + "pct_cuda_time": 0.046260624908481096, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.369, + "pct_cuda_time": 0.046260624908481096, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 224.415, + "cuda_time_us": 218.582, + "pct_cuda_time": 1.587649538977173, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 85.528, + "cuda_time_us": 126.638, + "pct_cuda_time": 0.9198230518386292, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT96x128x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT6_2_MO40_NTn1_NTA4_NTB0_NTC0_NTD4_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB2_WSGRA0_WSGRB0_WS64_WG16_16_1", + "cpu_time_us": 0, + "cuda_time_us": 126.638, + "pct_cuda_time": 0.9198230518386292, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 28672]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 28672]) <- linear(bfloat16[128, 4096], bfloat16[28672, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 26.84, + "cuda_time_us": 11.816, + "pct_cuda_time": 0.08582439062939436, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 11.816, + "pct_cuda_time": 0.08582439062939436, + "trace": "_C::silu_and_mul(bfloat16[128, 14336], bfloat16[128, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 94.271, + "cuda_time_us": 80.128, + "pct_cuda_time": 0.5820020965091495, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT128x64x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB1024_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_4_MO40_NTn1_NTA4_NTB0_NTC0_NTD0_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB4_WSGRA0_WSGRB0_WS64_WG64_4_1", + "cpu_time_us": 0, + "cuda_time_us": 75.522, + "pct_cuda_time": 0.5485468541903454, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + }, + { + "entry": { + "name": "Cijk_SB_BiasS_HAS_ScaleAlphaVec_PostGSU4_VW4", + "cpu_time_us": 0, + "cuda_time_us": 4.606, + "pct_cuda_time": 0.033455242318804194, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 1060.483, + "cuda_time_us": 423.538, + "pct_cuda_time": 3.076327924711614, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 37.136, + "cuda_time_us": 6.368, + "pct_cuda_time": 0.04625336150372235, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.368, + "pct_cuda_time": 0.04625336150372235, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 728.378, + "cuda_time_us": 184.58800000000002, + "pct_cuda_time": 1.340737357608213, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 85.898, + "cuda_time_us": 39.188, + "pct_cuda_time": 0.2846383056859095, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x48x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA2048_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT4_3_MO40_NTn1_NTA0_NTB0_NTC0_NTD0_NTM0_NEPBS8_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW4_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA4_VWB1_WSGRA0_WSGRB0_WS64_WG16_4_4", + "cpu_time_us": 0, + "cuda_time_us": 39.188, + "pct_cuda_time": 0.2846383056859095, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 6144]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 6144]) <- linear(bfloat16[128, 4096], bfloat16[6144, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 72.488, + "cuda_time_us": 20.947, + "pct_cuda_time": 0.15214653948154394, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.366, + "pct_cuda_time": 0.03171202517670411, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False) <- clone(bfloat16[128, 4096], 0) <- contiguous(bfloat16[128, 4096], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.925, + "pct_cuda_time": 0.028508863678095193, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False) <- clone(bfloat16[128, 1024], 0) <- contiguous(bfloat16[128, 1024], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.886, + "pct_cuda_time": 0.0354889956512543, + "trace": "_C::rotary_embedding(int64[128], bfloat16[128, 4096], bfloat16[128, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.406, + "pct_cuda_time": 0.03200256136705412, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.364, + "pct_cuda_time": 0.024434093608436236, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 429.25, + "cuda_time_us": 76.432, + "pct_cuda_time": 0.5551565525208082, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 4.526, + "pct_cuda_time": 0.032874169938104165, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[128], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "attn_fwd", + "cpu_time_us": 0, + "cuda_time_us": 71.906, + "pct_cuda_time": 0.522282382582704, + "trace": "_attention(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], int32[2], int32[2], 128, 128, True, 0.08838834764831845, None, None, None) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 102.143, + "cuda_time_us": 48.021, + "pct_cuda_time": 0.34879595991995144, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x32x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA1024_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_2_MO40_NTn1_NTA5_NTB0_NTC0_NTD7_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB1_WSGRA0_WSGRB0_WS64_WG32_4_2", + "cpu_time_us": 0, + "cuda_time_us": 48.021, + "pct_cuda_time": 0.34879595991995144, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 4096]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 4096]) <- linear(bfloat16[128, 4096], bfloat16[4096, 4096], None)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 38.718, + "cuda_time_us": 6.088, + "pct_cuda_time": 0.04421960817127224, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.088, + "pct_cuda_time": 0.04421960817127224, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 230.534, + "cuda_time_us": 226.494, + "pct_cuda_time": 1.6451175974284058, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 87.07, + "cuda_time_us": 134.84, + "pct_cuda_time": 0.9793974976698996, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT96x128x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT6_2_MO40_NTn1_NTA4_NTB0_NTC0_NTD4_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB2_WSGRA0_WSGRB0_WS64_WG16_16_1", + "cpu_time_us": 0, + "cuda_time_us": 134.84, + "pct_cuda_time": 0.9793974976698996, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 28672]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 28672]) <- linear(bfloat16[128, 4096], bfloat16[28672, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 28.803, + "cuda_time_us": 12.297, + "pct_cuda_time": 0.08931808831835329, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 12.297, + "pct_cuda_time": 0.08931808831835329, + "trace": "_C::silu_and_mul(bfloat16[128, 14336], bfloat16[128, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 96.955, + "cuda_time_us": 79.357, + "pct_cuda_time": 0.576402011440153, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT128x64x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB1024_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_4_MO40_NTn1_NTA4_NTB0_NTC0_NTD0_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB4_WSGRA0_WSGRB0_WS64_WG64_4_1", + "cpu_time_us": 0, + "cuda_time_us": 74.671, + "pct_cuda_time": 0.5423656967406488, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + }, + { + "entry": { + "name": "Cijk_SB_BiasS_HAS_ScaleAlphaVec_PostGSU4_VW4", + "cpu_time_us": 0, + "cuda_time_us": 4.686, + "pct_cuda_time": 0.03403631469950423, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 1034.425, + "cuda_time_us": 416.67600000000004, + "pct_cuda_time": 3.026486441257069, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 33.68, + "cuda_time_us": 6.288, + "pct_cuda_time": 0.04567228912302232, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.288, + "pct_cuda_time": 0.04567228912302232, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 720.225, + "cuda_time_us": 186.369, + "pct_cuda_time": 1.3536734814835474, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 84.016, + "cuda_time_us": 40.159, + "pct_cuda_time": 0.29169107170665604, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x48x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA2048_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT4_3_MO40_NTn1_NTA0_NTB0_NTC0_NTD0_NTM0_NEPBS8_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW4_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA4_VWB1_WSGRA0_WSGRB0_WS64_WG16_4_4", + "cpu_time_us": 0, + "cuda_time_us": 40.159, + "pct_cuda_time": 0.29169107170665604, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 6144]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 6144]) <- linear(bfloat16[128, 4096], bfloat16[6144, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 75.703, + "cuda_time_us": 20.665, + "pct_cuda_time": 0.15009825933957635, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.366, + "pct_cuda_time": 0.03171202517670411, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False) <- clone(bfloat16[128, 4096], 0) <- contiguous(bfloat16[128, 4096], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.884, + "pct_cuda_time": 0.028211064082986428, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False) <- clone(bfloat16[128, 1024], 0) <- contiguous(bfloat16[128, 1024], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 5.006, + "pct_cuda_time": 0.036360604222304346, + "trace": "_C::rotary_embedding(int64[128], bfloat16[128, 4096], bfloat16[128, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.885, + "pct_cuda_time": 0.02821832748774518, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.524, + "pct_cuda_time": 0.0255962383698363, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 425.986, + "cuda_time_us": 77.634, + "pct_cuda_time": 0.5638871650408261, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 5.287, + "pct_cuda_time": 0.038401620959513195, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[128], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "attn_fwd", + "cpu_time_us": 0, + "cuda_time_us": 72.347, + "pct_cuda_time": 0.5254855440813129, + "trace": "_attention(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], int32[2], int32[2], 128, 128, True, 0.08838834764831845, None, None, None) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 99.409, + "cuda_time_us": 47.911, + "pct_cuda_time": 0.34799698539648893, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x32x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA1024_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_2_MO40_NTn1_NTA5_NTB0_NTC0_NTD7_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB1_WSGRA0_WSGRB0_WS64_WG32_4_2", + "cpu_time_us": 0, + "cuda_time_us": 47.911, + "pct_cuda_time": 0.34799698539648893, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 4096]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 4096]) <- linear(bfloat16[128, 4096], bfloat16[4096, 4096], None)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 40.961, + "cuda_time_us": 6.288, + "pct_cuda_time": 0.04567228912302232, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.288, + "pct_cuda_time": 0.04567228912302232, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 219.538, + "cuda_time_us": 217.73100000000002, + "pct_cuda_time": 1.5814683815274768, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 84.236, + "cuda_time_us": 126.558, + "pct_cuda_time": 0.9192419794579292, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT96x128x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT6_2_MO40_NTn1_NTA4_NTB0_NTC0_NTD4_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB2_WSGRA0_WSGRB0_WS64_WG16_16_1", + "cpu_time_us": 0, + "cuda_time_us": 126.558, + "pct_cuda_time": 0.9192419794579292, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 28672]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 28672]) <- linear(bfloat16[128, 4096], bfloat16[28672, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 30.035, + "cuda_time_us": 12.017, + "pct_cuda_time": 0.08728433498590317, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 12.017, + "pct_cuda_time": 0.08728433498590317, + "trace": "_C::silu_and_mul(bfloat16[128, 14336], bfloat16[128, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 88.242, + "cuda_time_us": 79.156, + "pct_cuda_time": 0.5749420670836441, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT128x64x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB1024_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_4_MO40_NTn1_NTA4_NTB0_NTC0_NTD0_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB4_WSGRA0_WSGRB0_WS64_WG64_4_1", + "cpu_time_us": 0, + "cuda_time_us": 74.43, + "pct_cuda_time": 0.5406152161937899, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + }, + { + "entry": { + "name": "Cijk_SB_BiasS_HAS_ScaleAlphaVec_PostGSU4_VW4", + "cpu_time_us": 0, + "cuda_time_us": 4.726, + "pct_cuda_time": 0.034326850889854245, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 1069.747, + "cuda_time_us": 420.50300000000004, + "pct_cuda_time": 3.0542834912688064, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 33.941, + "cuda_time_us": 6.128, + "pct_cuda_time": 0.04451014436162226, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.128, + "pct_cuda_time": 0.04451014436162226, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 731.122, + "cuda_time_us": 188.40200000000002, + "pct_cuda_time": 1.3684399833580871, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 84.125, + "cuda_time_us": 40.369, + "pct_cuda_time": 0.29321638670599365, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x48x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA2048_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT4_3_MO40_NTn1_NTA0_NTB0_NTC0_NTD0_NTM0_NEPBS8_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW4_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA4_VWB1_WSGRA0_WSGRB0_WS64_WG16_4_4", + "cpu_time_us": 0, + "cuda_time_us": 40.369, + "pct_cuda_time": 0.29321638670599365, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 6144]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 6144]) <- linear(bfloat16[128, 4096], bfloat16[6144, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 72.438, + "cuda_time_us": 21.347, + "pct_cuda_time": 0.15505190138504413, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.285, + "pct_cuda_time": 0.031123689391245327, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False) <- clone(bfloat16[128, 4096], 0) <- contiguous(bfloat16[128, 4096], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.046, + "pct_cuda_time": 0.029387735653903995, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False) <- clone(bfloat16[128, 1024], 0) <- contiguous(bfloat16[128, 1024], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 5.006, + "pct_cuda_time": 0.036360604222304346, + "trace": "_C::rotary_embedding(int64[128], bfloat16[128, 4096], bfloat16[128, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.646, + "pct_cuda_time": 0.03374577850915421, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.364, + "pct_cuda_time": 0.024434093608436236, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 434.008, + "cuda_time_us": 76.47200000000001, + "pct_cuda_time": 0.5554470887111582, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 5.287, + "pct_cuda_time": 0.038401620959513195, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[128], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "attn_fwd", + "cpu_time_us": 0, + "cuda_time_us": 71.185, + "pct_cuda_time": 0.517045467751645, + "trace": "_attention(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], int32[2], int32[2], 128, 128, True, 0.08838834764831845, None, None, None) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 104.626, + "cuda_time_us": 50.214, + "pct_cuda_time": 0.364724606555891, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x32x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA1024_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_2_MO40_NTn1_NTA5_NTB0_NTC0_NTD7_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB1_WSGRA0_WSGRB0_WS64_WG32_4_2", + "cpu_time_us": 0, + "cuda_time_us": 50.214, + "pct_cuda_time": 0.364724606555891, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 4096]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 4096]) <- linear(bfloat16[128, 4096], bfloat16[4096, 4096], None)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 39.68, + "cuda_time_us": 6.088, + "pct_cuda_time": 0.04421960817127224, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.088, + "pct_cuda_time": 0.04421960817127224, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 242.752, + "cuda_time_us": 219.885, + "pct_cuda_time": 1.5971137553778245, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 99.629, + "cuda_time_us": 126.759, + "pct_cuda_time": 0.9207019238144379, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT96x128x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT6_2_MO40_NTn1_NTA4_NTB0_NTC0_NTD4_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB2_WSGRA0_WSGRB0_WS64_WG16_16_1", + "cpu_time_us": 0, + "cuda_time_us": 126.759, + "pct_cuda_time": 0.9207019238144379, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 28672]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 28672]) <- linear(bfloat16[128, 4096], bfloat16[28672, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 27.601, + "cuda_time_us": 12.016, + "pct_cuda_time": 0.08727707158114442, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 12.016, + "pct_cuda_time": 0.08727707158114442, + "trace": "_C::silu_and_mul(bfloat16[128, 14336], bfloat16[128, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 97.175, + "cuda_time_us": 81.11, + "pct_cuda_time": 0.5891347599822423, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT128x64x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB1024_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_4_MO40_NTn1_NTA4_NTB0_NTC0_NTD0_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB4_WSGRA0_WSGRB0_WS64_WG64_4_1", + "cpu_time_us": 0, + "cuda_time_us": 76.624, + "pct_cuda_time": 0.5565511262344882, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + }, + { + "entry": { + "name": "Cijk_SB_BiasS_HAS_ScaleAlphaVec_PostGSU4_VW4", + "cpu_time_us": 0, + "cuda_time_us": 4.486, + "pct_cuda_time": 0.03258363374775415, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 1053.322, + "cuda_time_us": 424.577, + "pct_cuda_time": 3.083874602255955, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 35.874, + "cuda_time_us": 6.168, + "pct_cuda_time": 0.044800680551972276, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.168, + "pct_cuda_time": 0.044800680551972276, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 728.036, + "cuda_time_us": 186.60799999999998, + "pct_cuda_time": 1.3554094352208885, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 86.709, + "cuda_time_us": 40.239, + "pct_cuda_time": 0.292272144087356, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x48x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA2048_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT4_3_MO40_NTn1_NTA0_NTB0_NTC0_NTD0_NTM0_NEPBS8_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW4_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA4_VWB1_WSGRA0_WSGRB0_WS64_WG16_4_4", + "cpu_time_us": 0, + "cuda_time_us": 40.239, + "pct_cuda_time": 0.292272144087356, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 6144]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 6144]) <- linear(bfloat16[128, 4096], bfloat16[6144, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 72.237, + "cuda_time_us": 20.505000000000003, + "pct_cuda_time": 0.14893611457817632, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.325, + "pct_cuda_time": 0.031414225581595345, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False) <- clone(bfloat16[128, 4096], 0) <- contiguous(bfloat16[128, 4096], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.885, + "pct_cuda_time": 0.02821832748774518, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False) <- clone(bfloat16[128, 1024], 0) <- contiguous(bfloat16[128, 1024], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.806, + "pct_cuda_time": 0.03490792327055427, + "trace": "_C::rotary_embedding(int64[128], bfloat16[128, 4096], bfloat16[128, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.165, + "pct_cuda_time": 0.030252080820195287, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.324, + "pct_cuda_time": 0.024143557418086225, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 433.156, + "cuda_time_us": 78.074, + "pct_cuda_time": 0.5670830631346763, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 5.487, + "pct_cuda_time": 0.03985430191126327, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[128], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "attn_fwd", + "cpu_time_us": 0, + "cuda_time_us": 72.587, + "pct_cuda_time": 0.527228761223413, + "trace": "_attention(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], int32[2], int32[2], 128, 128, True, 0.08838834764831845, None, None, None) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 98.838, + "cuda_time_us": 47.79, + "pct_cuda_time": 0.3471181134206801, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x32x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA1024_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_2_MO40_NTn1_NTA5_NTB0_NTC0_NTD7_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB1_WSGRA0_WSGRB0_WS64_WG32_4_2", + "cpu_time_us": 0, + "cuda_time_us": 47.79, + "pct_cuda_time": 0.3471181134206801, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 4096]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 4096]) <- linear(bfloat16[128, 4096], bfloat16[4096, 4096], None)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 38.998, + "cuda_time_us": 6.048, + "pct_cuda_time": 0.043929071980922225, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.048, + "pct_cuda_time": 0.043929071980922225, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 224.325, + "cuda_time_us": 225.75300000000001, + "pct_cuda_time": 1.639735414502172, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 87.14, + "cuda_time_us": 131.476, + "pct_cuda_time": 0.9549634040614634, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT96x128x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT6_2_MO40_NTn1_NTA4_NTB0_NTC0_NTD4_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB2_WSGRA0_WSGRB0_WS64_WG16_16_1", + "cpu_time_us": 0, + "cuda_time_us": 131.476, + "pct_cuda_time": 0.9549634040614634, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 28672]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 28672]) <- linear(bfloat16[128, 4096], bfloat16[28672, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 26.029, + "cuda_time_us": 12.257, + "pct_cuda_time": 0.08902755212800327, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 12.257, + "pct_cuda_time": 0.08902755212800327, + "trace": "_C::silu_and_mul(bfloat16[128, 14336], bfloat16[128, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 90.645, + "cuda_time_us": 82.02000000000001, + "pct_cuda_time": 0.5957444583127053, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT128x64x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB1024_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_4_MO40_NTn1_NTA4_NTB0_NTC0_NTD0_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB4_WSGRA0_WSGRB0_WS64_WG64_4_1", + "cpu_time_us": 0, + "cuda_time_us": 77.575, + "pct_cuda_time": 0.5634586241600599, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + }, + { + "entry": { + "name": "Cijk_SB_BiasS_HAS_ScaleAlphaVec_PostGSU4_VW4", + "cpu_time_us": 0, + "cuda_time_us": 4.445, + "pct_cuda_time": 0.03228583415264539, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 1052.02, + "cuda_time_us": 421.355, + "pct_cuda_time": 3.0604719121232615, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 36.014, + "cuda_time_us": 6.329, + "pct_cuda_time": 0.04597008871813108, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.329, + "pct_cuda_time": 0.04597008871813108, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 725.703, + "cuda_time_us": 184.45700000000002, + "pct_cuda_time": 1.339785851584817, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 84.176, + "cuda_time_us": 40.259, + "pct_cuda_time": 0.2924174121825311, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x48x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA2048_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT4_3_MO40_NTn1_NTA0_NTB0_NTC0_NTD0_NTM0_NEPBS8_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW4_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA4_VWB1_WSGRA0_WSGRB0_WS64_WG16_4_4", + "cpu_time_us": 0, + "cuda_time_us": 40.259, + "pct_cuda_time": 0.2924174121825311, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 6144]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 6144]) <- linear(bfloat16[128, 4096], bfloat16[6144, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 76.624, + "cuda_time_us": 20.907, + "pct_cuda_time": 0.15185600329119398, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.286, + "pct_cuda_time": 0.03113095279600408, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False) <- clone(bfloat16[128, 4096], 0) <- contiguous(bfloat16[128, 4096], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.765, + "pct_cuda_time": 0.027346718916695135, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False) <- clone(bfloat16[128, 1024], 0) <- contiguous(bfloat16[128, 1024], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.726, + "pct_cuda_time": 0.034326850889854245, + "trace": "_C::rotary_embedding(int64[128], bfloat16[128, 4096], bfloat16[128, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.366, + "pct_cuda_time": 0.03171202517670411, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.764, + "pct_cuda_time": 0.027339455511936388, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 425.856, + "cuda_time_us": 76.432, + "pct_cuda_time": 0.5551565525208082, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 5.286, + "pct_cuda_time": 0.03839435755475445, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[128], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "attn_fwd", + "cpu_time_us": 0, + "cuda_time_us": 71.146, + "pct_cuda_time": 0.5167621949660538, + "trace": "_attention(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], int32[2], int32[2], 128, 128, True, 0.08838834764831845, None, None, None) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 102.153, + "cuda_time_us": 46.859, + "pct_cuda_time": 0.3403558835902835, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x32x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA1024_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_2_MO40_NTn1_NTA5_NTB0_NTC0_NTD7_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB1_WSGRA0_WSGRB0_WS64_WG32_4_2", + "cpu_time_us": 0, + "cuda_time_us": 46.859, + "pct_cuda_time": 0.3403558835902835, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 4096]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 4096]) <- linear(bfloat16[128, 4096], bfloat16[4096, 4096], None)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 38.277, + "cuda_time_us": 6.448, + "pct_cuda_time": 0.04683443388442238, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.448, + "pct_cuda_time": 0.04683443388442238, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 226.158, + "cuda_time_us": 224.12099999999998, + "pct_cuda_time": 1.627881537935891, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 87.811, + "cuda_time_us": 132.968, + "pct_cuda_time": 0.9658004039615189, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT96x128x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT6_2_MO40_NTn1_NTA4_NTB0_NTC0_NTD4_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB2_WSGRA0_WSGRB0_WS64_WG16_16_1", + "cpu_time_us": 0, + "cuda_time_us": 132.968, + "pct_cuda_time": 0.9658004039615189, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 28672]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 28672]) <- linear(bfloat16[128, 4096], bfloat16[28672, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 26.29, + "cuda_time_us": 11.696, + "pct_cuda_time": 0.0849527820583443, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 11.696, + "pct_cuda_time": 0.0849527820583443, + "trace": "_C::silu_and_mul(bfloat16[128, 14336], bfloat16[128, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 94.631, + "cuda_time_us": 79.45700000000001, + "pct_cuda_time": 0.5771283519160281, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT128x64x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB1024_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_4_MO40_NTn1_NTA4_NTB0_NTC0_NTD0_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB4_WSGRA0_WSGRB0_WS64_WG64_4_1", + "cpu_time_us": 0, + "cuda_time_us": 74.891, + "pct_cuda_time": 0.5439636457875738, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + }, + { + "entry": { + "name": "Cijk_SB_BiasS_HAS_ScaleAlphaVec_PostGSU4_VW4", + "cpu_time_us": 0, + "cuda_time_us": 4.566, + "pct_cuda_time": 0.03316470612845418, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 1067.433, + "cuda_time_us": 421.264, + "pct_cuda_time": 3.0598109422902153, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 36.445, + "cuda_time_us": 6.248, + "pct_cuda_time": 0.045381752932672305, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.248, + "pct_cuda_time": 0.045381752932672305, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 744.282, + "cuda_time_us": 186.179, + "pct_cuda_time": 1.352293434579385, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 87.431, + "cuda_time_us": 40.339, + "pct_cuda_time": 0.29299848456323113, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x48x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA2048_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT4_3_MO40_NTn1_NTA0_NTB0_NTC0_NTD0_NTM0_NEPBS8_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW4_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA4_VWB1_WSGRA0_WSGRB0_WS64_WG16_4_4", + "cpu_time_us": 0, + "cuda_time_us": 40.339, + "pct_cuda_time": 0.29299848456323113, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 6144]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 6144]) <- linear(bfloat16[128, 4096], bfloat16[6144, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 75.994, + "cuda_time_us": 20.466, + "pct_cuda_time": 0.14865284179258503, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.205, + "pct_cuda_time": 0.030542617010545302, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False) <- clone(bfloat16[128, 4096], 0) <- contiguous(bfloat16[128, 4096], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.885, + "pct_cuda_time": 0.02821832748774518, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False) <- clone(bfloat16[128, 1024], 0) <- contiguous(bfloat16[128, 1024], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.766, + "pct_cuda_time": 0.03461738708020426, + "trace": "_C::rotary_embedding(int64[128], bfloat16[128, 4096], bfloat16[128, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.286, + "pct_cuda_time": 0.03113095279600408, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.324, + "pct_cuda_time": 0.024143557418086225, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 440.297, + "cuda_time_us": 77.714, + "pct_cuda_time": 0.5644682374215261, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 5.327, + "pct_cuda_time": 0.03869215714986322, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[128], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "attn_fwd", + "cpu_time_us": 0, + "cuda_time_us": 72.387, + "pct_cuda_time": 0.5257760802716629, + "trace": "_attention(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], int32[2], int32[2], 128, 128, True, 0.08838834764831845, None, None, None) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 103.384, + "cuda_time_us": 47.66, + "pct_cuda_time": 0.34617387080204254, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x32x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA1024_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_2_MO40_NTn1_NTA5_NTB0_NTC0_NTD7_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB1_WSGRA0_WSGRB0_WS64_WG32_4_2", + "cpu_time_us": 0, + "cuda_time_us": 47.66, + "pct_cuda_time": 0.34617387080204254, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 4096]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 4096]) <- linear(bfloat16[128, 4096], bfloat16[4096, 4096], None)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 42.293, + "cuda_time_us": 5.968, + "pct_cuda_time": 0.0433479996002222, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 5.968, + "pct_cuda_time": 0.0433479996002222, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 223.103, + "cuda_time_us": 222.86900000000003, + "pct_cuda_time": 1.6187877551779362, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 84.796, + "cuda_time_us": 127.67, + "pct_cuda_time": 0.9273188855496596, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT96x128x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT6_2_MO40_NTn1_NTA4_NTB0_NTC0_NTD4_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB2_WSGRA0_WSGRB0_WS64_WG16_16_1", + "cpu_time_us": 0, + "cuda_time_us": 127.67, + "pct_cuda_time": 0.9273188855496596, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 28672]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 28672]) <- linear(bfloat16[128, 4096], bfloat16[28672, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 28.222, + "cuda_time_us": 12.458, + "pct_cuda_time": 0.0904874964845121, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 12.458, + "pct_cuda_time": 0.0904874964845121, + "trace": "_C::silu_and_mul(bfloat16[128, 14336], bfloat16[128, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 93.619, + "cuda_time_us": 82.741, + "pct_cuda_time": 0.6009813731437642, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT128x64x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB1024_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_4_MO40_NTn1_NTA4_NTB0_NTC0_NTD0_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB4_WSGRA0_WSGRB0_WS64_WG64_4_1", + "cpu_time_us": 0, + "cuda_time_us": 78.215, + "pct_cuda_time": 0.5681072032056601, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + }, + { + "entry": { + "name": "Cijk_SB_BiasS_HAS_ScaleAlphaVec_PostGSU4_VW4", + "cpu_time_us": 0, + "cuda_time_us": 4.526, + "pct_cuda_time": 0.032874169938104165, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 1060.784, + "cuda_time_us": 427.40399999999994, + "pct_cuda_time": 3.104408247508942, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 35.984, + "cuda_time_us": 6.128, + "pct_cuda_time": 0.04451014436162226, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.128, + "pct_cuda_time": 0.04451014436162226, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 733.195, + "cuda_time_us": 187.28199999999998, + "pct_cuda_time": 1.3603049700282863, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 82.573, + "cuda_time_us": 40.199, + "pct_cuda_time": 0.29198160789700606, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x48x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA2048_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT4_3_MO40_NTn1_NTA0_NTB0_NTC0_NTD0_NTM0_NEPBS8_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW4_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA4_VWB1_WSGRA0_WSGRB0_WS64_WG16_4_4", + "cpu_time_us": 0, + "cuda_time_us": 40.199, + "pct_cuda_time": 0.29198160789700606, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 6144]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 6144]) <- linear(bfloat16[128, 4096], bfloat16[6144, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 79.679, + "cuda_time_us": 21.548000000000002, + "pct_cuda_time": 0.15651184574155297, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.366, + "pct_cuda_time": 0.03171202517670411, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False) <- clone(bfloat16[128, 4096], 0) <- contiguous(bfloat16[128, 4096], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.965, + "pct_cuda_time": 0.028799399868445208, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False) <- clone(bfloat16[128, 1024], 0) <- contiguous(bfloat16[128, 1024], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.927, + "pct_cuda_time": 0.035786795246363065, + "trace": "_C::rotary_embedding(int64[128], bfloat16[128, 4096], bfloat16[128, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.405, + "pct_cuda_time": 0.031995297962295374, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.885, + "pct_cuda_time": 0.02821832748774518, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 433.017, + "cuda_time_us": 76.833, + "pct_cuda_time": 0.5580691778290671, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 5.567, + "pct_cuda_time": 0.040435374291963304, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[128], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "attn_fwd", + "cpu_time_us": 0, + "cuda_time_us": 71.266, + "pct_cuda_time": 0.5176338035371038, + "trace": "_attention(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], int32[2], int32[2], 128, 128, True, 0.08838834764831845, None, None, None) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 101.922, + "cuda_time_us": 48.702, + "pct_cuda_time": 0.35374233856066045, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x32x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA1024_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_2_MO40_NTn1_NTA5_NTB0_NTC0_NTD7_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB1_WSGRA0_WSGRB0_WS64_WG32_4_2", + "cpu_time_us": 0, + "cuda_time_us": 48.702, + "pct_cuda_time": 0.35374233856066045, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 4096]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 4096]) <- linear(bfloat16[128, 4096], bfloat16[4096, 4096], None)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 40.14, + "cuda_time_us": 6.368, + "pct_cuda_time": 0.04625336150372235, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.368, + "pct_cuda_time": 0.04625336150372235, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 229.162, + "cuda_time_us": 227.62599999999998, + "pct_cuda_time": 1.653339771615311, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 88.582, + "cuda_time_us": 133.809, + "pct_cuda_time": 0.9719089273636281, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT96x128x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT6_2_MO40_NTn1_NTA4_NTB0_NTC0_NTD4_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB2_WSGRA0_WSGRB0_WS64_WG16_16_1", + "cpu_time_us": 0, + "cuda_time_us": 133.809, + "pct_cuda_time": 0.9719089273636281, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 28672]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 28672]) <- linear(bfloat16[128, 4096], bfloat16[28672, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 28.723, + "cuda_time_us": 12.898, + "pct_cuda_time": 0.09368339457836225, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 12.898, + "pct_cuda_time": 0.09368339457836225, + "trace": "_C::silu_and_mul(bfloat16[128, 14336], bfloat16[128, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 94.781, + "cuda_time_us": 80.919, + "pct_cuda_time": 0.587747449673321, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT128x64x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB1024_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_4_MO40_NTn1_NTA4_NTB0_NTC0_NTD0_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB4_WSGRA0_WSGRB0_WS64_WG64_4_1", + "cpu_time_us": 0, + "cuda_time_us": 76.473, + "pct_cuda_time": 0.5554543521159169, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + }, + { + "entry": { + "name": "Cijk_SB_BiasS_HAS_ScaleAlphaVec_PostGSU4_VW4", + "cpu_time_us": 0, + "cuda_time_us": 4.446, + "pct_cuda_time": 0.032293097557404137, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 1038.4, + "cuda_time_us": 418.4, + "pct_cuda_time": 3.039008551061154, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 32.879, + "cuda_time_us": 6.208, + "pct_cuda_time": 0.04509121674232229, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.208, + "pct_cuda_time": 0.04509121674232229, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 721.557, + "cuda_time_us": 184.858, + "pct_cuda_time": 1.3426984768930756, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 82.413, + "cuda_time_us": 39.759, + "pct_cuda_time": 0.28878570980315593, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x48x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA2048_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT4_3_MO40_NTn1_NTA0_NTB0_NTC0_NTD0_NTM0_NEPBS8_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW4_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA4_VWB1_WSGRA0_WSGRB0_WS64_WG16_4_4", + "cpu_time_us": 0, + "cuda_time_us": 39.759, + "pct_cuda_time": 0.28878570980315593, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 6144]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 6144]) <- linear(bfloat16[128, 4096], bfloat16[6144, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 73.279, + "cuda_time_us": 20.906, + "pct_cuda_time": 0.15184873988643519, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.446, + "pct_cuda_time": 0.032293097557404137, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False) <- clone(bfloat16[128, 4096], 0) <- contiguous(bfloat16[128, 4096], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.925, + "pct_cuda_time": 0.028508863678095193, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False) <- clone(bfloat16[128, 1024], 0) <- contiguous(bfloat16[128, 1024], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 5.086, + "pct_cuda_time": 0.036941676603004375, + "trace": "_C::rotary_embedding(int64[128], bfloat16[128, 4096], bfloat16[128, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.885, + "pct_cuda_time": 0.02821832748774518, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.564, + "pct_cuda_time": 0.025886774560186315, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 430.102, + "cuda_time_us": 76.39200000000001, + "pct_cuda_time": 0.5548660163304582, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 5.367, + "pct_cuda_time": 0.03898269334021323, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[128], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "attn_fwd", + "cpu_time_us": 0, + "cuda_time_us": 71.025, + "pct_cuda_time": 0.515883322990245, + "trace": "_attention(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], int32[2], int32[2], 128, 128, True, 0.08838834764831845, None, None, None) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 98.818, + "cuda_time_us": 47.801, + "pct_cuda_time": 0.3471980108730264, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x32x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA1024_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_2_MO40_NTn1_NTA5_NTB0_NTC0_NTD7_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB1_WSGRA0_WSGRB0_WS64_WG32_4_2", + "cpu_time_us": 0, + "cuda_time_us": 47.801, + "pct_cuda_time": 0.3471980108730264, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 4096]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 4096]) <- linear(bfloat16[128, 4096], bfloat16[4096, 4096], None)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 38.467, + "cuda_time_us": 5.967, + "pct_cuda_time": 0.04334073619546345, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 5.967, + "pct_cuda_time": 0.04334073619546345, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 219.758, + "cuda_time_us": 221.36699999999996, + "pct_cuda_time": 1.6078781212302926, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 86.188, + "cuda_time_us": 129.082, + "pct_cuda_time": 0.9375748130690149, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT96x128x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT6_2_MO40_NTn1_NTA4_NTB0_NTC0_NTD4_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB2_WSGRA0_WSGRB0_WS64_WG16_16_1", + "cpu_time_us": 0, + "cuda_time_us": 129.082, + "pct_cuda_time": 0.9375748130690149, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 28672]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 28672]) <- linear(bfloat16[128, 4096], bfloat16[28672, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 25.598, + "cuda_time_us": 12.658, + "pct_cuda_time": 0.09194017743626216, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 12.658, + "pct_cuda_time": 0.09194017743626216, + "trace": "_C::silu_and_mul(bfloat16[128, 14336], bfloat16[128, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 90.405, + "cuda_time_us": 79.627, + "pct_cuda_time": 0.5783631307250156, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT128x64x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB1024_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_4_MO40_NTn1_NTA4_NTB0_NTC0_NTD0_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB4_WSGRA0_WSGRB0_WS64_WG64_4_1", + "cpu_time_us": 0, + "cuda_time_us": 75.261, + "pct_cuda_time": 0.5466511055483114, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + }, + { + "entry": { + "name": "Cijk_SB_BiasS_HAS_ScaleAlphaVec_PostGSU4_VW4", + "cpu_time_us": 0, + "cuda_time_us": 4.366, + "pct_cuda_time": 0.03171202517670411, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 1043.949, + "cuda_time_us": 421.855, + "pct_cuda_time": 3.064103614502637, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 33.94, + "cuda_time_us": 6.288, + "pct_cuda_time": 0.04567228912302232, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.288, + "pct_cuda_time": 0.04567228912302232, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 726.074, + "cuda_time_us": 186.981, + "pct_cuda_time": 1.3581186851959026, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 83.325, + "cuda_time_us": 39.748, + "pct_cuda_time": 0.2887058123508096, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x48x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA2048_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT4_3_MO40_NTn1_NTA0_NTB0_NTC0_NTD0_NTM0_NEPBS8_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW4_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA4_VWB1_WSGRA0_WSGRB0_WS64_WG16_4_4", + "cpu_time_us": 0, + "cuda_time_us": 39.748, + "pct_cuda_time": 0.2887058123508096, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 6144]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 6144]) <- linear(bfloat16[128, 4096], bfloat16[6144, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 75.383, + "cuda_time_us": 21.427999999999997, + "pct_cuda_time": 0.15564023717050288, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.406, + "pct_cuda_time": 0.03200256136705412, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False) <- clone(bfloat16[128, 4096], 0) <- contiguous(bfloat16[128, 4096], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.045, + "pct_cuda_time": 0.02938047224914524, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False) <- clone(bfloat16[128, 1024], 0) <- contiguous(bfloat16[128, 1024], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.967, + "pct_cuda_time": 0.03607733143671308, + "trace": "_C::rotary_embedding(int64[128], bfloat16[128, 4096], bfloat16[128, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.566, + "pct_cuda_time": 0.03316470612845418, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.444, + "pct_cuda_time": 0.02501516598913627, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 432.676, + "cuda_time_us": 77.354, + "pct_cuda_time": 0.561853411708376, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 5.608, + "pct_cuda_time": 0.040733173887072066, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[128], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "attn_fwd", + "cpu_time_us": 0, + "cuda_time_us": 71.746, + "pct_cuda_time": 0.5211202378213039, + "trace": "_attention(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], int32[2], int32[2], 128, 128, True, 0.08838834764831845, None, None, None) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 99.168, + "cuda_time_us": 48.451, + "pct_cuda_time": 0.3519192239662141, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x32x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA1024_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_2_MO40_NTn1_NTA5_NTB0_NTC0_NTD7_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB1_WSGRA0_WSGRB0_WS64_WG32_4_2", + "cpu_time_us": 0, + "cuda_time_us": 48.451, + "pct_cuda_time": 0.3519192239662141, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 4096]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 4096]) <- linear(bfloat16[128, 4096], bfloat16[4096, 4096], None)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 45.938, + "cuda_time_us": 6.288, + "pct_cuda_time": 0.04567228912302232, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.288, + "pct_cuda_time": 0.04567228912302232, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 217.775, + "cuda_time_us": 222.29799999999997, + "pct_cuda_time": 1.6146403510606895, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 85.568, + "cuda_time_us": 129.623, + "pct_cuda_time": 0.9415043150434989, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT96x128x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT6_2_MO40_NTn1_NTA4_NTB0_NTC0_NTD4_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB2_WSGRA0_WSGRB0_WS64_WG16_16_1", + "cpu_time_us": 0, + "cuda_time_us": 129.623, + "pct_cuda_time": 0.9415043150434989, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 28672]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 28672]) <- linear(bfloat16[128, 4096], bfloat16[28672, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 25.969, + "cuda_time_us": 12.057, + "pct_cuda_time": 0.08757487117625319, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 12.057, + "pct_cuda_time": 0.08757487117625319, + "trace": "_C::silu_and_mul(bfloat16[128, 14336], bfloat16[128, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 88.442, + "cuda_time_us": 80.618, + "pct_cuda_time": 0.5855611648409371, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT128x64x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB1024_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_4_MO40_NTn1_NTA4_NTB0_NTC0_NTD0_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB4_WSGRA0_WSGRB0_WS64_WG64_4_1", + "cpu_time_us": 0, + "cuda_time_us": 76.092, + "pct_cuda_time": 0.552686994902833, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + }, + { + "entry": { + "name": "Cijk_SB_BiasS_HAS_ScaleAlphaVec_PostGSU4_VW4", + "cpu_time_us": 0, + "cuda_time_us": 4.526, + "pct_cuda_time": 0.032874169938104165, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 1036.247, + "cuda_time_us": 422.393, + "pct_cuda_time": 3.068011326262844, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 32.659, + "cuda_time_us": 6.289, + "pct_cuda_time": 0.04567955252778106, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.289, + "pct_cuda_time": 0.04567955252778106, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 722.008, + "cuda_time_us": 181.89, + "pct_cuda_time": 1.3211406915691044, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 84.335, + "cuda_time_us": 40.79, + "pct_cuda_time": 0.2962742801094275, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x48x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA2048_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT4_3_MO40_NTn1_NTA0_NTB0_NTC0_NTD0_NTM0_NEPBS8_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW4_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA4_VWB1_WSGRA0_WSGRB0_WS64_WG16_4_4", + "cpu_time_us": 0, + "cuda_time_us": 40.79, + "pct_cuda_time": 0.2962742801094275, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 6144]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 6144]) <- linear(bfloat16[128, 4096], bfloat16[6144, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 72.448, + "cuda_time_us": 21.265, + "pct_cuda_time": 0.1544563021948266, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.486, + "pct_cuda_time": 0.03258363374775415, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False) <- clone(bfloat16[128, 4096], 0) <- contiguous(bfloat16[128, 4096], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.045, + "pct_cuda_time": 0.02938047224914524, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False) <- clone(bfloat16[128, 1024], 0) <- contiguous(bfloat16[128, 1024], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 5.006, + "pct_cuda_time": 0.036360604222304346, + "trace": "_C::rotary_embedding(int64[128], bfloat16[128, 4096], bfloat16[128, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.445, + "pct_cuda_time": 0.03228583415264539, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.283, + "pct_cuda_time": 0.02384575782297746, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 419.997, + "cuda_time_us": 74.789, + "pct_cuda_time": 0.5432227785021813, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 5.527, + "pct_cuda_time": 0.04014483810161329, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[128], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "attn_fwd", + "cpu_time_us": 0, + "cuda_time_us": 69.262, + "pct_cuda_time": 0.5030779404005681, + "trace": "_attention(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], int32[2], int32[2], 128, 128, True, 0.08838834764831845, None, None, None) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 104.326, + "cuda_time_us": 45.046, + "pct_cuda_time": 0.3271873307626691, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x32x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA1024_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_2_MO40_NTn1_NTA5_NTB0_NTC0_NTD7_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB1_WSGRA0_WSGRB0_WS64_WG32_4_2", + "cpu_time_us": 0, + "cuda_time_us": 45.046, + "pct_cuda_time": 0.3271873307626691, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 4096]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 4096]) <- linear(bfloat16[128, 4096], bfloat16[4096, 4096], None)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 39.198, + "cuda_time_us": 6.769, + "pct_cuda_time": 0.04916598681198125, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.769, + "pct_cuda_time": 0.04916598681198125, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 221.771, + "cuda_time_us": 227.445, + "pct_cuda_time": 1.6520250953539775, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 85.547, + "cuda_time_us": 136.573, + "pct_cuda_time": 0.9919849781168142, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT96x128x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT6_2_MO40_NTn1_NTA4_NTB0_NTC0_NTD4_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB2_WSGRA0_WSGRB0_WS64_WG16_16_1", + "cpu_time_us": 0, + "cuda_time_us": 136.573, + "pct_cuda_time": 0.9919849781168142, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 28672]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 28672]) <- linear(bfloat16[128, 4096], bfloat16[28672, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 26.619, + "cuda_time_us": 12.177, + "pct_cuda_time": 0.08844647974730323, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 12.177, + "pct_cuda_time": 0.08844647974730323, + "trace": "_C::silu_and_mul(bfloat16[128, 14336], bfloat16[128, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 92.579, + "cuda_time_us": 78.69500000000001, + "pct_cuda_time": 0.5715936374898603, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT128x64x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB1024_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_4_MO40_NTn1_NTA4_NTB0_NTC0_NTD0_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB4_WSGRA0_WSGRB0_WS64_WG64_4_1", + "cpu_time_us": 0, + "cuda_time_us": 74.09, + "pct_cuda_time": 0.5381456585758149, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + }, + { + "entry": { + "name": "Cijk_SB_BiasS_HAS_ScaleAlphaVec_PostGSU4_VW4", + "cpu_time_us": 0, + "cuda_time_us": 4.605, + "pct_cuda_time": 0.03344797891404545, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 1038.46, + "cuda_time_us": 421.624, + "pct_cuda_time": 3.0624257680033655, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 35.513, + "cuda_time_us": 6.369, + "pct_cuda_time": 0.046260624908481096, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.369, + "pct_cuda_time": 0.046260624908481096, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 732.013, + "cuda_time_us": 186.22799999999998, + "pct_cuda_time": 1.3526493414125635, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 83.204, + "cuda_time_us": 40.72, + "pct_cuda_time": 0.295765841776315, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x48x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA2048_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT4_3_MO40_NTn1_NTA0_NTB0_NTC0_NTD0_NTM0_NEPBS8_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW4_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA4_VWB1_WSGRA0_WSGRB0_WS64_WG16_4_4", + "cpu_time_us": 0, + "cuda_time_us": 40.72, + "pct_cuda_time": 0.295765841776315, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 6144]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 6144]) <- linear(bfloat16[128, 4096], bfloat16[6144, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 73.479, + "cuda_time_us": 21.026, + "pct_cuda_time": 0.15272034845748525, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.326, + "pct_cuda_time": 0.03142148898635409, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False) <- clone(bfloat16[128, 4096], 0) <- contiguous(bfloat16[128, 4096], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.965, + "pct_cuda_time": 0.028799399868445208, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False) <- clone(bfloat16[128, 1024], 0) <- contiguous(bfloat16[128, 1024], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.886, + "pct_cuda_time": 0.0354889956512543, + "trace": "_C::rotary_embedding(int64[128], bfloat16[128, 4096], bfloat16[128, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.325, + "pct_cuda_time": 0.031414225581595345, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.524, + "pct_cuda_time": 0.0255962383698363, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 439.206, + "cuda_time_us": 78.03399999999999, + "pct_cuda_time": 0.5667925269443261, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 5.487, + "pct_cuda_time": 0.03985430191126327, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[128], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "attn_fwd", + "cpu_time_us": 0, + "cuda_time_us": 72.547, + "pct_cuda_time": 0.526938225033063, + "trace": "_attention(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], int32[2], int32[2], 128, 128, True, 0.08838834764831845, None, None, None) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 100.48, + "cuda_time_us": 46.448, + "pct_cuda_time": 0.33737062423443714, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x32x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA1024_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_2_MO40_NTn1_NTA5_NTB0_NTC0_NTD7_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB1_WSGRA0_WSGRB0_WS64_WG32_4_2", + "cpu_time_us": 0, + "cuda_time_us": 46.448, + "pct_cuda_time": 0.33737062423443714, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 4096]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 4096]) <- linear(bfloat16[128, 4096], bfloat16[4096, 4096], None)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 38.117, + "cuda_time_us": 6.369, + "pct_cuda_time": 0.046260624908481096, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.369, + "pct_cuda_time": 0.046260624908481096, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 212.036, + "cuda_time_us": 222.65800000000002, + "pct_cuda_time": 1.6172551767738397, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 82.563, + "cuda_time_us": 127.91, + "pct_cuda_time": 0.9290621026917596, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT96x128x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT6_2_MO40_NTn1_NTA4_NTB0_NTC0_NTD4_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB2_WSGRA0_WSGRB0_WS64_WG16_16_1", + "cpu_time_us": 0, + "cuda_time_us": 127.91, + "pct_cuda_time": 0.9290621026917596, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 28672]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 28672]) <- linear(bfloat16[128, 4096], bfloat16[28672, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 25.989, + "cuda_time_us": 12.698, + "pct_cuda_time": 0.09223071362661218, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 12.698, + "pct_cuda_time": 0.09223071362661218, + "trace": "_C::silu_and_mul(bfloat16[128, 14336], bfloat16[128, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 86.239, + "cuda_time_us": 82.05000000000001, + "pct_cuda_time": 0.5959623604554678, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT128x64x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB1024_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_4_MO40_NTn1_NTA4_NTB0_NTC0_NTD0_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB4_WSGRA0_WSGRB0_WS64_WG64_4_1", + "cpu_time_us": 0, + "cuda_time_us": 77.644, + "pct_cuda_time": 0.5639597990884136, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + }, + { + "entry": { + "name": "Cijk_SB_BiasS_HAS_ScaleAlphaVec_PostGSU4_VW4", + "cpu_time_us": 0, + "cuda_time_us": 4.406, + "pct_cuda_time": 0.03200256136705412, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 1016.357, + "cuda_time_us": 422.447, + "pct_cuda_time": 3.068403550119817, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 33.37, + "cuda_time_us": 6.409, + "pct_cuda_time": 0.04655116109883111, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.409, + "pct_cuda_time": 0.04655116109883111, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 709.339, + "cuda_time_us": 185.298, + "pct_cuda_time": 1.3458943749869259, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 83.235, + "cuda_time_us": 39.197, + "pct_cuda_time": 0.28470367632873816, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x48x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA2048_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT4_3_MO40_NTn1_NTA0_NTB0_NTC0_NTD0_NTM0_NEPBS8_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW4_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA4_VWB1_WSGRA0_WSGRB0_WS64_WG16_4_4", + "cpu_time_us": 0, + "cuda_time_us": 39.197, + "pct_cuda_time": 0.28470367632873816, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 6144]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 6144]) <- linear(bfloat16[128, 4096], bfloat16[6144, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 71.386, + "cuda_time_us": 20.585, + "pct_cuda_time": 0.14951718695887634, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.205, + "pct_cuda_time": 0.030542617010545302, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False) <- clone(bfloat16[128, 4096], 0) <- contiguous(bfloat16[128, 4096], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.925, + "pct_cuda_time": 0.028508863678095193, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False) <- clone(bfloat16[128, 1024], 0) <- contiguous(bfloat16[128, 1024], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 5.006, + "pct_cuda_time": 0.036360604222304346, + "trace": "_C::rotary_embedding(int64[128], bfloat16[128, 4096], bfloat16[128, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.125, + "pct_cuda_time": 0.029961544629845273, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.324, + "pct_cuda_time": 0.024143557418086225, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 421.139, + "cuda_time_us": 75.19200000000001, + "pct_cuda_time": 0.5461499306199578, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 5.408, + "pct_cuda_time": 0.039280492935321994, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[128], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "attn_fwd", + "cpu_time_us": 0, + "cuda_time_us": 69.784, + "pct_cuda_time": 0.5068694376846358, + "trace": "_attention(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], int32[2], int32[2], 128, 128, True, 0.08838834764831845, None, None, None) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 98.658, + "cuda_time_us": 50.324, + "pct_cuda_time": 0.3655235810793535, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x32x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA1024_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_2_MO40_NTn1_NTA5_NTB0_NTC0_NTD7_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB1_WSGRA0_WSGRB0_WS64_WG32_4_2", + "cpu_time_us": 0, + "cuda_time_us": 50.324, + "pct_cuda_time": 0.3655235810793535, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 4096]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 4096]) <- linear(bfloat16[128, 4096], bfloat16[4096, 4096], None)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 36.835, + "cuda_time_us": 6.008, + "pct_cuda_time": 0.04363853579057221, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.008, + "pct_cuda_time": 0.04363853579057221, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 216.212, + "cuda_time_us": 224.73199999999997, + "pct_cuda_time": 1.6323194782434876, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 84.546, + "cuda_time_us": 132.457, + "pct_cuda_time": 0.9620888041297975, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT96x128x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT6_2_MO40_NTn1_NTA4_NTB0_NTC0_NTD4_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB2_WSGRA0_WSGRB0_WS64_WG16_16_1", + "cpu_time_us": 0, + "cuda_time_us": 132.457, + "pct_cuda_time": 0.9620888041297975, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 28672]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 28672]) <- linear(bfloat16[128, 4096], bfloat16[28672, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 25.077, + "cuda_time_us": 11.736, + "pct_cuda_time": 0.08524331824869433, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 11.736, + "pct_cuda_time": 0.08524331824869433, + "trace": "_C::silu_and_mul(bfloat16[128, 14336], bfloat16[128, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 88.542, + "cuda_time_us": 80.539, + "pct_cuda_time": 0.584987355864996, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT128x64x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB1024_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_4_MO40_NTn1_NTA4_NTB0_NTC0_NTD0_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB4_WSGRA0_WSGRB0_WS64_WG64_4_1", + "cpu_time_us": 0, + "cuda_time_us": 76.173, + "pct_cuda_time": 0.5532753306882918, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + }, + { + "entry": { + "name": "Cijk_SB_BiasS_HAS_ScaleAlphaVec_PostGSU4_VW4", + "cpu_time_us": 0, + "cuda_time_us": 4.366, + "pct_cuda_time": 0.03171202517670411, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 1021.855, + "cuda_time_us": 422.135, + "pct_cuda_time": 3.0661373678350867, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 32.699, + "cuda_time_us": 6.288, + "pct_cuda_time": 0.04567228912302232, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.288, + "pct_cuda_time": 0.04567228912302232, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 715.678, + "cuda_time_us": 184.396, + "pct_cuda_time": 1.3393427838945327, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 84.827, + "cuda_time_us": 39.568, + "pct_cuda_time": 0.28739839949423457, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x48x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA2048_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT4_3_MO40_NTn1_NTA0_NTB0_NTC0_NTD0_NTM0_NEPBS8_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW4_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA4_VWB1_WSGRA0_WSGRB0_WS64_WG16_4_4", + "cpu_time_us": 0, + "cuda_time_us": 39.568, + "pct_cuda_time": 0.28739839949423457, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 6144]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 6144]) <- linear(bfloat16[128, 4096], bfloat16[6144, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 71.096, + "cuda_time_us": 19.785, + "pct_cuda_time": 0.14370646315187602, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.206, + "pct_cuda_time": 0.03054988041530405, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False) <- clone(bfloat16[128, 4096], 0) <- contiguous(bfloat16[128, 4096], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.885, + "pct_cuda_time": 0.02821832748774518, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False) <- clone(bfloat16[128, 1024], 0) <- contiguous(bfloat16[128, 1024], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.726, + "pct_cuda_time": 0.034326850889854245, + "trace": "_C::rotary_embedding(int64[128], bfloat16[128, 4096], bfloat16[128, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.684, + "pct_cuda_time": 0.026758383131236355, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.284, + "pct_cuda_time": 0.02385302122773621, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 423.412, + "cuda_time_us": 77.914, + "pct_cuda_time": 0.5659209183732762, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 5.167, + "pct_cuda_time": 0.03753001238846315, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[128], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "attn_fwd", + "cpu_time_us": 0, + "cuda_time_us": 72.747, + "pct_cuda_time": 0.5283909059848131, + "trace": "_attention(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], int32[2], int32[2], 128, 128, True, 0.08838834764831845, None, None, None) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 97.676, + "cuda_time_us": 47.129, + "pct_cuda_time": 0.3423170028751461, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x32x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA1024_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_2_MO40_NTn1_NTA5_NTB0_NTC0_NTD7_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB1_WSGRA0_WSGRB0_WS64_WG32_4_2", + "cpu_time_us": 0, + "cuda_time_us": 47.129, + "pct_cuda_time": 0.3423170028751461, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 4096]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 4096]) <- linear(bfloat16[128, 4096], bfloat16[4096, 4096], None)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 38.418, + "cuda_time_us": 6.449, + "pct_cuda_time": 0.046841697289181125, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.449, + "pct_cuda_time": 0.046841697289181125, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 214.54, + "cuda_time_us": 225.00199999999998, + "pct_cuda_time": 1.63428059752835, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 84.726, + "cuda_time_us": 133.438, + "pct_cuda_time": 0.9692142041981315, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT96x128x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT6_2_MO40_NTn1_NTA4_NTB0_NTC0_NTD4_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB2_WSGRA0_WSGRB0_WS64_WG16_16_1", + "cpu_time_us": 0, + "cuda_time_us": 133.438, + "pct_cuda_time": 0.9692142041981315, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 28672]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 28672]) <- linear(bfloat16[128, 4096], bfloat16[28672, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 25.168, + "cuda_time_us": 11.897, + "pct_cuda_time": 0.08641272641485313, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 11.897, + "pct_cuda_time": 0.08641272641485313, + "trace": "_C::silu_and_mul(bfloat16[128, 14336], bfloat16[128, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 86.94, + "cuda_time_us": 79.667, + "pct_cuda_time": 0.5786536669153656, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT128x64x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB1024_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_4_MO40_NTn1_NTA4_NTB0_NTC0_NTD0_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB4_WSGRA0_WSGRB0_WS64_WG64_4_1", + "cpu_time_us": 0, + "cuda_time_us": 75.141, + "pct_cuda_time": 0.5457794969772615, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + }, + { + "entry": { + "name": "Cijk_SB_BiasS_HAS_ScaleAlphaVec_PostGSU4_VW4", + "cpu_time_us": 0, + "cuda_time_us": 4.526, + "pct_cuda_time": 0.032874169938104165, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 1014.675, + "cuda_time_us": 419.392, + "pct_cuda_time": 3.0462138485818344, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 34.782, + "cuda_time_us": 6.368, + "pct_cuda_time": 0.04625336150372235, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.368, + "pct_cuda_time": 0.04625336150372235, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 706.715, + "cuda_time_us": 187.853, + "pct_cuda_time": 1.364452374145533, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 80.931, + "cuda_time_us": 41.06, + "pct_cuda_time": 0.29823539939429017, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x48x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA2048_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT4_3_MO40_NTn1_NTA0_NTB0_NTC0_NTD0_NTM0_NEPBS8_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW4_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA4_VWB1_WSGRA0_WSGRB0_WS64_WG16_4_4", + "cpu_time_us": 0, + "cuda_time_us": 41.06, + "pct_cuda_time": 0.29823539939429017, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 6144]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 6144]) <- linear(bfloat16[128, 4096], bfloat16[6144, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 73.489, + "cuda_time_us": 20.387, + "pct_cuda_time": 0.14807903281664375, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.965, + "pct_cuda_time": 0.028799399868445208, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False) <- clone(bfloat16[128, 4096], 0) <- contiguous(bfloat16[128, 4096], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.805, + "pct_cuda_time": 0.027637255107045153, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False) <- clone(bfloat16[128, 1024], 0) <- contiguous(bfloat16[128, 1024], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.847, + "pct_cuda_time": 0.035205722865663036, + "trace": "_C::rotary_embedding(int64[128], bfloat16[128, 4096], bfloat16[128, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.286, + "pct_cuda_time": 0.03113095279600408, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.484, + "pct_cuda_time": 0.025305702179486283, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 418.184, + "cuda_time_us": 75.831, + "pct_cuda_time": 0.5507912462607992, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 5.407, + "pct_cuda_time": 0.03927322953056324, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[128], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "attn_fwd", + "cpu_time_us": 0, + "cuda_time_us": 70.424, + "pct_cuda_time": 0.511518016730236, + "trace": "_attention(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], int32[2], int32[2], 128, 128, True, 0.08838834764831845, None, None, None) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 98.678, + "cuda_time_us": 50.575, + "pct_cuda_time": 0.3673466956737999, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x32x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA1024_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_2_MO40_NTn1_NTA5_NTB0_NTC0_NTD7_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB1_WSGRA0_WSGRB0_WS64_WG32_4_2", + "cpu_time_us": 0, + "cuda_time_us": 50.575, + "pct_cuda_time": 0.3673466956737999, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 4096]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 4096]) <- linear(bfloat16[128, 4096], bfloat16[4096, 4096], None)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 37.776, + "cuda_time_us": 6.048, + "pct_cuda_time": 0.043929071980922225, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.048, + "pct_cuda_time": 0.043929071980922225, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 214.451, + "cuda_time_us": 219.123, + "pct_cuda_time": 1.5915790409516568, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 82.664, + "cuda_time_us": 126.588, + "pct_cuda_time": 0.9194598816006917, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT96x128x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT6_2_MO40_NTn1_NTA4_NTB0_NTC0_NTD4_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB2_WSGRA0_WSGRB0_WS64_WG16_16_1", + "cpu_time_us": 0, + "cuda_time_us": 126.588, + "pct_cuda_time": 0.9194598816006917, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 28672]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 28672]) <- linear(bfloat16[128, 4096], bfloat16[28672, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 25.107, + "cuda_time_us": 12.257, + "pct_cuda_time": 0.08902755212800327, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 12.257, + "pct_cuda_time": 0.08902755212800327, + "trace": "_C::silu_and_mul(bfloat16[128, 14336], bfloat16[128, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 88.773, + "cuda_time_us": 80.278, + "pct_cuda_time": 0.583091607222962, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT128x64x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB1024_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_4_MO40_NTn1_NTA4_NTB0_NTC0_NTD0_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB4_WSGRA0_WSGRB0_WS64_WG64_4_1", + "cpu_time_us": 0, + "cuda_time_us": 75.872, + "pct_cuda_time": 0.551089045855908, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + }, + { + "entry": { + "name": "Cijk_SB_BiasS_HAS_ScaleAlphaVec_PostGSU4_VW4", + "cpu_time_us": 0, + "cuda_time_us": 4.406, + "pct_cuda_time": 0.03200256136705412, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 1039.141, + "cuda_time_us": 426.812, + "pct_cuda_time": 3.1001083118917623, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 33.5, + "cuda_time_us": 6.048, + "pct_cuda_time": 0.043929071980922225, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.048, + "pct_cuda_time": 0.043929071980922225, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 726.955, + "cuda_time_us": 185.398, + "pct_cuda_time": 1.3466207154628007, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 85.047, + "cuda_time_us": 38.176, + "pct_cuda_time": 0.2772877400700541, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x48x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA2048_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT4_3_MO40_NTn1_NTA0_NTB0_NTC0_NTD0_NTM0_NEPBS8_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW4_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA4_VWB1_WSGRA0_WSGRB0_WS64_WG16_4_4", + "cpu_time_us": 0, + "cuda_time_us": 38.176, + "pct_cuda_time": 0.2772877400700541, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 6144]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 6144]) <- linear(bfloat16[128, 4096], bfloat16[6144, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 71.216, + "cuda_time_us": 21.186, + "pct_cuda_time": 0.1538824932188853, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.366, + "pct_cuda_time": 0.03171202517670411, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False) <- clone(bfloat16[128, 4096], 0) <- contiguous(bfloat16[128, 4096], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.085, + "pct_cuda_time": 0.029671008439495255, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False) <- clone(bfloat16[128, 1024], 0) <- contiguous(bfloat16[128, 1024], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.966, + "pct_cuda_time": 0.03607006803195433, + "trace": "_C::rotary_embedding(int64[128], bfloat16[128, 4096], bfloat16[128, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.325, + "pct_cuda_time": 0.031414225581595345, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.444, + "pct_cuda_time": 0.02501516598913627, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 432.897, + "cuda_time_us": 77.55399999999999, + "pct_cuda_time": 0.563306092660126, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 5.487, + "pct_cuda_time": 0.03985430191126327, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[128], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "attn_fwd", + "cpu_time_us": 0, + "cuda_time_us": 72.067, + "pct_cuda_time": 0.5234517907488628, + "trace": "_attention(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], int32[2], int32[2], 128, 128, True, 0.08838834764831845, None, None, None) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 97.085, + "cuda_time_us": 48.482, + "pct_cuda_time": 0.35214438951373533, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x32x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA1024_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_2_MO40_NTn1_NTA5_NTB0_NTC0_NTD7_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB1_WSGRA0_WSGRB0_WS64_WG32_4_2", + "cpu_time_us": 0, + "cuda_time_us": 48.482, + "pct_cuda_time": 0.35214438951373533, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 4096]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 4096]) <- linear(bfloat16[128, 4096], bfloat16[4096, 4096], None)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 37.396, + "cuda_time_us": 6.328, + "pct_cuda_time": 0.045962825313372334, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.328, + "pct_cuda_time": 0.045962825313372334, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 214.761, + "cuda_time_us": 229.038, + "pct_cuda_time": 1.6635956991346668, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 83.514, + "cuda_time_us": 135.361, + "pct_cuda_time": 0.9831817315492085, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT96x128x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT6_2_MO40_NTn1_NTA4_NTB0_NTC0_NTD4_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB2_WSGRA0_WSGRB0_WS64_WG16_16_1", + "cpu_time_us": 0, + "cuda_time_us": 135.361, + "pct_cuda_time": 0.9831817315492085, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 28672]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 28672]) <- linear(bfloat16[128, 4096], bfloat16[28672, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 25.829, + "cuda_time_us": 13.019, + "pct_cuda_time": 0.09456226655417105, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 13.019, + "pct_cuda_time": 0.09456226655417105, + "trace": "_C::silu_and_mul(bfloat16[128, 14336], bfloat16[128, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 88.262, + "cuda_time_us": 80.658, + "pct_cuda_time": 0.5858517010312873, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT128x64x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB1024_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_4_MO40_NTn1_NTA4_NTB0_NTC0_NTD0_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB4_WSGRA0_WSGRB0_WS64_WG64_4_1", + "cpu_time_us": 0, + "cuda_time_us": 76.092, + "pct_cuda_time": 0.552686994902833, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + }, + { + "entry": { + "name": "Cijk_SB_BiasS_HAS_ScaleAlphaVec_PostGSU4_VW4", + "cpu_time_us": 0, + "cuda_time_us": 4.566, + "pct_cuda_time": 0.03316470612845418, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 1033.082, + "cuda_time_us": 417.82, + "pct_cuda_time": 3.034795776301079, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 35.132, + "cuda_time_us": 6.489, + "pct_cuda_time": 0.04713223347953114, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.489, + "pct_cuda_time": 0.04713223347953114, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 716.149, + "cuda_time_us": 184.337, + "pct_cuda_time": 1.3389142430137666, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 84.556, + "cuda_time_us": 38.677, + "pct_cuda_time": 0.280926705854188, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x48x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA2048_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT4_3_MO40_NTn1_NTA0_NTB0_NTC0_NTD0_NTM0_NEPBS8_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW4_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA4_VWB1_WSGRA0_WSGRB0_WS64_WG16_4_4", + "cpu_time_us": 0, + "cuda_time_us": 38.677, + "pct_cuda_time": 0.280926705854188, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 6144]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 6144]) <- linear(bfloat16[128, 4096], bfloat16[6144, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 70.405, + "cuda_time_us": 20.785999999999998, + "pct_cuda_time": 0.15097713131538515, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.486, + "pct_cuda_time": 0.03258363374775415, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False) <- clone(bfloat16[128, 4096], 0) <- contiguous(bfloat16[128, 4096], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.085, + "pct_cuda_time": 0.029671008439495255, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False) <- clone(bfloat16[128, 1024], 0) <- contiguous(bfloat16[128, 1024], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 5.086, + "pct_cuda_time": 0.036941676603004375, + "trace": "_C::rotary_embedding(int64[128], bfloat16[128, 4096], bfloat16[128, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.845, + "pct_cuda_time": 0.027927791297395168, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.284, + "pct_cuda_time": 0.02385302122773621, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 418.605, + "cuda_time_us": 76.673, + "pct_cuda_time": 0.556907033067667, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 5.648, + "pct_cuda_time": 0.041023710077422074, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[128], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "attn_fwd", + "cpu_time_us": 0, + "cuda_time_us": 71.025, + "pct_cuda_time": 0.515883322990245, + "trace": "_attention(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], int32[2], int32[2], 128, 128, True, 0.08838834764831845, None, None, None) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 99.028, + "cuda_time_us": 48.201, + "pct_cuda_time": 0.35010337277652653, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x32x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA1024_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_2_MO40_NTn1_NTA5_NTB0_NTC0_NTD7_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB1_WSGRA0_WSGRB0_WS64_WG32_4_2", + "cpu_time_us": 0, + "cuda_time_us": 48.201, + "pct_cuda_time": 0.35010337277652653, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 4096]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 4096]) <- linear(bfloat16[128, 4096], bfloat16[4096, 4096], None)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 36.655, + "cuda_time_us": 6.529, + "pct_cuda_time": 0.047422769669881154, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.529, + "pct_cuda_time": 0.047422769669881154, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 220.419, + "cuda_time_us": 220.465, + "pct_cuda_time": 1.6013265301379, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 86.039, + "cuda_time_us": 128.982, + "pct_cuda_time": 0.93684847259314, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT96x128x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT6_2_MO40_NTn1_NTA4_NTB0_NTC0_NTD4_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB2_WSGRA0_WSGRB0_WS64_WG16_16_1", + "cpu_time_us": 0, + "cuda_time_us": 128.982, + "pct_cuda_time": 0.93684847259314, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 28672]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 28672]) <- linear(bfloat16[128, 4096], bfloat16[28672, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 26.78, + "cuda_time_us": 11.936, + "pct_cuda_time": 0.0866959992004444, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 11.936, + "pct_cuda_time": 0.0866959992004444, + "trace": "_C::silu_and_mul(bfloat16[128, 14336], bfloat16[128, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 89.163, + "cuda_time_us": 79.547, + "pct_cuda_time": 0.5777820583443155, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT128x64x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB1024_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_4_MO40_NTn1_NTA4_NTB0_NTC0_NTD0_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB4_WSGRA0_WSGRB0_WS64_WG64_4_1", + "cpu_time_us": 0, + "cuda_time_us": 75.142, + "pct_cuda_time": 0.5457867603820201, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + }, + { + "entry": { + "name": "Cijk_SB_BiasS_HAS_ScaleAlphaVec_PostGSU4_VW4", + "cpu_time_us": 0, + "cuda_time_us": 4.405, + "pct_cuda_time": 0.031995297962295374, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 1008.836, + "cuda_time_us": 423.837, + "pct_cuda_time": 3.07849968273448, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 32.148, + "cuda_time_us": 6.369, + "pct_cuda_time": 0.046260624908481096, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.369, + "pct_cuda_time": 0.046260624908481096, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 703.801, + "cuda_time_us": 183.303, + "pct_cuda_time": 1.3314038824932186, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 82.082, + "cuda_time_us": 38.456, + "pct_cuda_time": 0.2793214934025042, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x48x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA2048_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT4_3_MO40_NTn1_NTA0_NTB0_NTC0_NTD0_NTM0_NEPBS8_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW4_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA4_VWB1_WSGRA0_WSGRB0_WS64_WG16_4_4", + "cpu_time_us": 0, + "cuda_time_us": 38.456, + "pct_cuda_time": 0.2793214934025042, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 6144]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 6144]) <- linear(bfloat16[128, 4096], bfloat16[6144, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 71.136, + "cuda_time_us": 20.505000000000003, + "pct_cuda_time": 0.14893611457817632, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.325, + "pct_cuda_time": 0.031414225581595345, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False) <- clone(bfloat16[128, 4096], 0) <- contiguous(bfloat16[128, 4096], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.924, + "pct_cuda_time": 0.028501600273336446, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False) <- clone(bfloat16[128, 1024], 0) <- contiguous(bfloat16[128, 1024], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.767, + "pct_cuda_time": 0.03462465048496301, + "trace": "_C::rotary_embedding(int64[128], bfloat16[128, 4096], bfloat16[128, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.125, + "pct_cuda_time": 0.029961544629845273, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.364, + "pct_cuda_time": 0.024434093608436236, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 418.445, + "cuda_time_us": 77.794, + "pct_cuda_time": 0.5650493098022261, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 5.327, + "pct_cuda_time": 0.03869215714986322, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[128], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "attn_fwd", + "cpu_time_us": 0, + "cuda_time_us": 72.467, + "pct_cuda_time": 0.5263571526523629, + "trace": "_attention(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], int32[2], int32[2], 128, 128, True, 0.08838834764831845, None, None, None) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 96.924, + "cuda_time_us": 46.548, + "pct_cuda_time": 0.33809696471031214, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x32x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA1024_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_2_MO40_NTn1_NTA5_NTB0_NTC0_NTD7_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB1_WSGRA0_WSGRB0_WS64_WG32_4_2", + "cpu_time_us": 0, + "cuda_time_us": 46.548, + "pct_cuda_time": 0.33809696471031214, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 4096]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 4096]) <- linear(bfloat16[128, 4096], bfloat16[4096, 4096], None)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 38.107, + "cuda_time_us": 6.288, + "pct_cuda_time": 0.04567228912302232, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.288, + "pct_cuda_time": 0.04567228912302232, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 213.799, + "cuda_time_us": 227.877, + "pct_cuda_time": 1.6551628862097576, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 83.274, + "cuda_time_us": 130.955, + "pct_cuda_time": 0.9511791701821546, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT96x128x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT6_2_MO40_NTn1_NTA4_NTB0_NTC0_NTD4_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB2_WSGRA0_WSGRB0_WS64_WG16_16_1", + "cpu_time_us": 0, + "cuda_time_us": 130.955, + "pct_cuda_time": 0.9511791701821546, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 28672]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 28672]) <- linear(bfloat16[128, 4096], bfloat16[28672, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 25.709, + "cuda_time_us": 12.017, + "pct_cuda_time": 0.08728433498590317, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 12.017, + "pct_cuda_time": 0.08728433498590317, + "trace": "_C::silu_and_mul(bfloat16[128, 14336], bfloat16[128, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 87.14, + "cuda_time_us": 84.905, + "pct_cuda_time": 0.6166993810417001, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT128x64x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB1024_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_4_MO40_NTn1_NTA4_NTB0_NTC0_NTD0_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB4_WSGRA0_WSGRB0_WS64_WG64_4_1", + "cpu_time_us": 0, + "cuda_time_us": 80.379, + "pct_cuda_time": 0.5838252111035959, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + }, + { + "entry": { + "name": "Cijk_SB_BiasS_HAS_ScaleAlphaVec_PostGSU4_VW4", + "cpu_time_us": 0, + "cuda_time_us": 4.526, + "pct_cuda_time": 0.032874169938104165, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 1039.832, + "cuda_time_us": 415.58500000000004, + "pct_cuda_time": 3.018562066665272, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 35.092, + "cuda_time_us": 6.328, + "pct_cuda_time": 0.045962825313372334, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.328, + "pct_cuda_time": 0.045962825313372334, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 725.543, + "cuda_time_us": 173.80100000000002, + "pct_cuda_time": 1.262387010475573, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 88.302, + "cuda_time_us": 38.527, + "pct_cuda_time": 0.2798371951403754, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x48x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA2048_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT4_3_MO40_NTn1_NTA0_NTB0_NTC0_NTD0_NTM0_NEPBS8_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW4_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA4_VWB1_WSGRA0_WSGRB0_WS64_WG16_4_4", + "cpu_time_us": 0, + "cuda_time_us": 38.527, + "pct_cuda_time": 0.2798371951403754, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 6144]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 6144]) <- linear(bfloat16[128, 4096], bfloat16[6144, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 71.116, + "cuda_time_us": 19.985, + "pct_cuda_time": 0.1451591441036261, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.125, + "pct_cuda_time": 0.029961544629845273, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False) <- clone(bfloat16[128, 4096], 0) <- contiguous(bfloat16[128, 4096], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.765, + "pct_cuda_time": 0.027346718916695135, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False) <- clone(bfloat16[128, 1024], 0) <- contiguous(bfloat16[128, 1024], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.846, + "pct_cuda_time": 0.03519845946090428, + "trace": "_C::rotary_embedding(int64[128], bfloat16[128, 4096], bfloat16[128, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.925, + "pct_cuda_time": 0.028508863678095193, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.324, + "pct_cuda_time": 0.024143557418086225, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 428.599, + "cuda_time_us": 76.432, + "pct_cuda_time": 0.5551565525208082, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 5.447, + "pct_cuda_time": 0.03956376572091325, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[128], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "attn_fwd", + "cpu_time_us": 0, + "cuda_time_us": 70.985, + "pct_cuda_time": 0.515592786799895, + "trace": "_attention(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], int32[2], int32[2], 128, 128, True, 0.08838834764831845, None, None, None) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 102.163, + "cuda_time_us": 38.857, + "pct_cuda_time": 0.28223411871076304, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x32x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA1024_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_2_MO40_NTn1_NTA5_NTB0_NTC0_NTD7_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB1_WSGRA0_WSGRB0_WS64_WG32_4_2", + "cpu_time_us": 0, + "cuda_time_us": 38.857, + "pct_cuda_time": 0.28223411871076304, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 4096]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 4096]) <- linear(bfloat16[128, 4096], bfloat16[4096, 4096], None)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 38.387, + "cuda_time_us": 6.168, + "pct_cuda_time": 0.044800680551972276, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.168, + "pct_cuda_time": 0.044800680551972276, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 219.167, + "cuda_time_us": 229.28799999999998, + "pct_cuda_time": 1.6654115503243545, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 86.158, + "cuda_time_us": 137.765, + "pct_cuda_time": 1.0006429565892445, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT96x128x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT6_2_MO40_NTn1_NTA4_NTB0_NTC0_NTD4_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB2_WSGRA0_WSGRB0_WS64_WG16_16_1", + "cpu_time_us": 0, + "cuda_time_us": 137.765, + "pct_cuda_time": 1.0006429565892445, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 28672]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 28672]) <- linear(bfloat16[128, 4096], bfloat16[28672, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 26.229, + "cuda_time_us": 12.658, + "pct_cuda_time": 0.09194017743626216, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 12.658, + "pct_cuda_time": 0.09194017743626216, + "trace": "_C::silu_and_mul(bfloat16[128, 14336], bfloat16[128, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 89.283, + "cuda_time_us": 78.86500000000001, + "pct_cuda_time": 0.5728284162988478, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT128x64x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB1024_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_4_MO40_NTn1_NTA4_NTB0_NTC0_NTD0_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB4_WSGRA0_WSGRB0_WS64_WG64_4_1", + "cpu_time_us": 0, + "cuda_time_us": 74.42, + "pct_cuda_time": 0.5405425821462024, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + }, + { + "entry": { + "name": "Cijk_SB_BiasS_HAS_ScaleAlphaVec_PostGSU4_VW4", + "cpu_time_us": 0, + "cuda_time_us": 4.445, + "pct_cuda_time": 0.03228583415264539, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 1055.105, + "cuda_time_us": 419.4409999999999, + "pct_cuda_time": 3.046569755415013, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 35.222, + "cuda_time_us": 6.128, + "pct_cuda_time": 0.04451014436162226, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.128, + "pct_cuda_time": 0.04451014436162226, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 739.414, + "cuda_time_us": 184.30599999999998, + "pct_cuda_time": 1.3386890774662452, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 86.789, + "cuda_time_us": 40.289, + "pct_cuda_time": 0.2926353143252936, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x48x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA2048_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT4_3_MO40_NTn1_NTA0_NTB0_NTC0_NTD0_NTM0_NEPBS8_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW4_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA4_VWB1_WSGRA0_WSGRB0_WS64_WG16_4_4", + "cpu_time_us": 0, + "cuda_time_us": 40.289, + "pct_cuda_time": 0.2926353143252936, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 6144]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 6144]) <- linear(bfloat16[128, 4096], bfloat16[6144, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 82.433, + "cuda_time_us": 20.906999999999996, + "pct_cuda_time": 0.15185600329119392, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.406, + "pct_cuda_time": 0.03200256136705412, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False) <- clone(bfloat16[128, 4096], 0) <- contiguous(bfloat16[128, 4096], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.925, + "pct_cuda_time": 0.028508863678095193, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False) <- clone(bfloat16[128, 1024], 0) <- contiguous(bfloat16[128, 1024], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.886, + "pct_cuda_time": 0.0354889956512543, + "trace": "_C::rotary_embedding(int64[128], bfloat16[128, 4096], bfloat16[128, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.366, + "pct_cuda_time": 0.03171202517670411, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.324, + "pct_cuda_time": 0.024143557418086225, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 429.611, + "cuda_time_us": 74.869, + "pct_cuda_time": 0.5438038508828813, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 5.407, + "pct_cuda_time": 0.03927322953056324, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[128], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "attn_fwd", + "cpu_time_us": 0, + "cuda_time_us": 69.462, + "pct_cuda_time": 0.5045306213523181, + "trace": "_attention(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], int32[2], int32[2], 128, 128, True, 0.08838834764831845, None, None, None) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 99.919, + "cuda_time_us": 48.241, + "pct_cuda_time": 0.3503939089668765, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x32x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA1024_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_2_MO40_NTn1_NTA5_NTB0_NTC0_NTD7_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB1_WSGRA0_WSGRB0_WS64_WG32_4_2", + "cpu_time_us": 0, + "cuda_time_us": 48.241, + "pct_cuda_time": 0.3503939089668765, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 4096]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 4096]) <- linear(bfloat16[128, 4096], bfloat16[4096, 4096], None)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 38.197, + "cuda_time_us": 6.088, + "pct_cuda_time": 0.04421960817127224, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.088, + "pct_cuda_time": 0.04421960817127224, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 219.758, + "cuda_time_us": 222.91899999999998, + "pct_cuda_time": 1.6191509254158731, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 84.826, + "cuda_time_us": 128.441, + "pct_cuda_time": 0.932918970618656, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT96x128x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT6_2_MO40_NTn1_NTA4_NTB0_NTC0_NTD4_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB2_WSGRA0_WSGRB0_WS64_WG16_16_1", + "cpu_time_us": 0, + "cuda_time_us": 128.441, + "pct_cuda_time": 0.932918970618656, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 28672]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 28672]) <- linear(bfloat16[128, 4096], bfloat16[28672, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 25.508, + "cuda_time_us": 12.858, + "pct_cuda_time": 0.09339285838801224, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 12.858, + "pct_cuda_time": 0.09339285838801224, + "trace": "_C::silu_and_mul(bfloat16[128, 14336], bfloat16[128, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 87.801, + "cuda_time_us": 81.61999999999999, + "pct_cuda_time": 0.592839096409205, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT128x64x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB1024_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_4_MO40_NTn1_NTA4_NTB0_NTC0_NTD0_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB4_WSGRA0_WSGRB0_WS64_WG64_4_1", + "cpu_time_us": 0, + "cuda_time_us": 77.014, + "pct_cuda_time": 0.5593838540904008, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + }, + { + "entry": { + "name": "Cijk_SB_BiasS_HAS_ScaleAlphaVec_PostGSU4_VW4", + "cpu_time_us": 0, + "cuda_time_us": 4.606, + "pct_cuda_time": 0.033455242318804194, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 1028.366, + "cuda_time_us": 411.086, + "pct_cuda_time": 2.985884008655654, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 33.259, + "cuda_time_us": 6.088, + "pct_cuda_time": 0.04421960817127224, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.088, + "pct_cuda_time": 0.04421960817127224, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 713.675, + "cuda_time_us": 173.89800000000002, + "pct_cuda_time": 1.2630915607371718, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 81.622, + "cuda_time_us": 37.665, + "pct_cuda_time": 0.2735761402383326, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x48x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA2048_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT4_3_MO40_NTn1_NTA0_NTB0_NTC0_NTD0_NTM0_NEPBS8_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW4_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA4_VWB1_WSGRA0_WSGRB0_WS64_WG16_4_4", + "cpu_time_us": 0, + "cuda_time_us": 37.665, + "pct_cuda_time": 0.2735761402383326, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 6144]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 6144]) <- linear(bfloat16[128, 4096], bfloat16[6144, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 71.576, + "cuda_time_us": 20.744000000000003, + "pct_cuda_time": 0.1506720683155177, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.245, + "pct_cuda_time": 0.030833153200895313, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False) <- clone(bfloat16[128, 4096], 0) <- contiguous(bfloat16[128, 4096], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.884, + "pct_cuda_time": 0.028211064082986428, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False) <- clone(bfloat16[128, 1024], 0) <- contiguous(bfloat16[128, 1024], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.886, + "pct_cuda_time": 0.0354889956512543, + "trace": "_C::rotary_embedding(int64[128], bfloat16[128, 4096], bfloat16[128, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.365, + "pct_cuda_time": 0.03170476177194536, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.364, + "pct_cuda_time": 0.024434093608436236, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 427.157, + "cuda_time_us": 77.15299999999999, + "pct_cuda_time": 0.5603934673518671, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 5.407, + "pct_cuda_time": 0.03927322953056324, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[128], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "attn_fwd", + "cpu_time_us": 0, + "cuda_time_us": 71.746, + "pct_cuda_time": 0.5211202378213039, + "trace": "_attention(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], int32[2], int32[2], 128, 128, True, 0.08838834764831845, None, None, None) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 96.885, + "cuda_time_us": 38.336, + "pct_cuda_time": 0.2784498848314541, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x32x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA1024_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_2_MO40_NTn1_NTA5_NTB0_NTC0_NTD7_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB1_WSGRA0_WSGRB0_WS64_WG32_4_2", + "cpu_time_us": 0, + "cuda_time_us": 38.336, + "pct_cuda_time": 0.2784498848314541, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 4096]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 4096]) <- linear(bfloat16[128, 4096], bfloat16[4096, 4096], None)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 37.816, + "cuda_time_us": 6.569, + "pct_cuda_time": 0.047713305860231175, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.569, + "pct_cuda_time": 0.047713305860231175, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 222.502, + "cuda_time_us": 224.531, + "pct_cuda_time": 1.630859533886979, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 83.605, + "cuda_time_us": 130.964, + "pct_cuda_time": 0.9512445408249833, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT96x128x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT6_2_MO40_NTn1_NTA4_NTB0_NTC0_NTD4_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB2_WSGRA0_WSGRB0_WS64_WG16_16_1", + "cpu_time_us": 0, + "cuda_time_us": 130.964, + "pct_cuda_time": 0.9512445408249833, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 28672]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 28672]) <- linear(bfloat16[128, 4096], bfloat16[28672, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 26.429, + "cuda_time_us": 12.257, + "pct_cuda_time": 0.08902755212800327, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 12.257, + "pct_cuda_time": 0.08902755212800327, + "trace": "_C::silu_and_mul(bfloat16[128, 14336], bfloat16[128, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 95.062, + "cuda_time_us": 81.31, + "pct_cuda_time": 0.5905874409339925, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT128x64x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB1024_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_4_MO40_NTn1_NTA4_NTB0_NTC0_NTD0_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB4_WSGRA0_WSGRB0_WS64_WG64_4_1", + "cpu_time_us": 0, + "cuda_time_us": 76.824, + "pct_cuda_time": 0.5580038071862383, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + }, + { + "entry": { + "name": "Cijk_SB_BiasS_HAS_ScaleAlphaVec_PostGSU4_VW4", + "cpu_time_us": 0, + "cuda_time_us": 4.486, + "pct_cuda_time": 0.03258363374775415, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 1038.39, + "cuda_time_us": 412.27200000000005, + "pct_cuda_time": 2.9944984066995324, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 35.413, + "cuda_time_us": 6.128, + "pct_cuda_time": 0.04451014436162226, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.128, + "pct_cuda_time": 0.04451014436162226, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 717.351, + "cuda_time_us": 176.246, + "pct_cuda_time": 1.2801460351107175, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 84.356, + "cuda_time_us": 40.049, + "pct_cuda_time": 0.2908920971831935, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x48x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA2048_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT4_3_MO40_NTn1_NTA0_NTB0_NTC0_NTD0_NTM0_NEPBS8_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW4_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA4_VWB1_WSGRA0_WSGRB0_WS64_WG16_4_4", + "cpu_time_us": 0, + "cuda_time_us": 40.049, + "pct_cuda_time": 0.2908920971831935, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 6144]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 6144]) <- linear(bfloat16[128, 4096], bfloat16[6144, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 70.996, + "cuda_time_us": 20.866999999999997, + "pct_cuda_time": 0.1515654671008439, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.406, + "pct_cuda_time": 0.03200256136705412, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False) <- clone(bfloat16[128, 4096], 0) <- contiguous(bfloat16[128, 4096], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.925, + "pct_cuda_time": 0.028508863678095193, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False) <- clone(bfloat16[128, 1024], 0) <- contiguous(bfloat16[128, 1024], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 5.046, + "pct_cuda_time": 0.03665114041265436, + "trace": "_C::rotary_embedding(int64[128], bfloat16[128, 4096], bfloat16[128, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.045, + "pct_cuda_time": 0.02938047224914524, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.445, + "pct_cuda_time": 0.025022429393895016, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 426.757, + "cuda_time_us": 77.675, + "pct_cuda_time": 0.5641849646359349, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 5.568, + "pct_cuda_time": 0.04044263769672205, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[128], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "attn_fwd", + "cpu_time_us": 0, + "cuda_time_us": 72.107, + "pct_cuda_time": 0.5237423269392129, + "trace": "_attention(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], int32[2], int32[2], 128, 128, True, 0.08838834764831845, None, None, None) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 100.44, + "cuda_time_us": 37.655, + "pct_cuda_time": 0.2735035061907451, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x32x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA1024_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_2_MO40_NTn1_NTA5_NTB0_NTC0_NTD7_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB1_WSGRA0_WSGRB0_WS64_WG32_4_2", + "cpu_time_us": 0, + "cuda_time_us": 37.655, + "pct_cuda_time": 0.2735035061907451, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 4096]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 4096]) <- linear(bfloat16[128, 4096], bfloat16[4096, 4096], None)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 38.367, + "cuda_time_us": 6.409, + "pct_cuda_time": 0.04655116109883111, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.409, + "pct_cuda_time": 0.04655116109883111, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 222.552, + "cuda_time_us": 223.489, + "pct_cuda_time": 1.623291066128361, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 86.68, + "cuda_time_us": 132.817, + "pct_cuda_time": 0.9647036298429477, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT96x128x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT6_2_MO40_NTn1_NTA4_NTB0_NTC0_NTD4_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB2_WSGRA0_WSGRB0_WS64_WG16_16_1", + "cpu_time_us": 0, + "cuda_time_us": 132.817, + "pct_cuda_time": 0.9647036298429477, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 28672]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 28672]) <- linear(bfloat16[128, 4096], bfloat16[28672, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 25.849, + "cuda_time_us": 12.337, + "pct_cuda_time": 0.0896086245087033, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 12.337, + "pct_cuda_time": 0.0896086245087033, + "trace": "_C::silu_and_mul(bfloat16[128, 14336], bfloat16[128, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 91.967, + "cuda_time_us": 78.33500000000001, + "pct_cuda_time": 0.5689788117767102, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT128x64x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB1024_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_4_MO40_NTn1_NTA4_NTB0_NTC0_NTD0_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB4_WSGRA0_WSGRB0_WS64_WG64_4_1", + "cpu_time_us": 0, + "cuda_time_us": 73.769, + "pct_cuda_time": 0.5358141056482559, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + }, + { + "entry": { + "name": "Cijk_SB_BiasS_HAS_ScaleAlphaVec_PostGSU4_VW4", + "cpu_time_us": 0, + "cuda_time_us": 4.566, + "pct_cuda_time": 0.03316470612845418, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 1036.737, + "cuda_time_us": 412.592, + "pct_cuda_time": 2.996822696222332, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 35.774, + "cuda_time_us": 6.85, + "pct_cuda_time": 0.049754322597440025, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.85, + "pct_cuda_time": 0.049754322597440025, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 720.906, + "cuda_time_us": 179.589, + "pct_cuda_time": 1.3044275972192199, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 84.547, + "cuda_time_us": 40.96, + "pct_cuda_time": 0.29750905891841506, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x48x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA2048_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT4_3_MO40_NTn1_NTA0_NTB0_NTC0_NTD0_NTM0_NEPBS8_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW4_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA4_VWB1_WSGRA0_WSGRB0_WS64_WG16_4_4", + "cpu_time_us": 0, + "cuda_time_us": 40.96, + "pct_cuda_time": 0.29750905891841506, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 6144]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 6144]) <- linear(bfloat16[128, 4096], bfloat16[6144, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 71.317, + "cuda_time_us": 20.826, + "pct_cuda_time": 0.15126766750573517, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.445, + "pct_cuda_time": 0.03228583415264539, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False) <- clone(bfloat16[128, 4096], 0) <- contiguous(bfloat16[128, 4096], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.285, + "pct_cuda_time": 0.031123689391245327, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False) <- clone(bfloat16[128, 1024], 0) <- contiguous(bfloat16[128, 1024], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 5.047, + "pct_cuda_time": 0.0366584038174131, + "trace": "_C::rotary_embedding(int64[128], bfloat16[128, 4096], bfloat16[128, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.805, + "pct_cuda_time": 0.027637255107045153, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.244, + "pct_cuda_time": 0.023562485037386196, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 431.765, + "cuda_time_us": 77.19300000000001, + "pct_cuda_time": 0.5606840035422173, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 5.647, + "pct_cuda_time": 0.04101644667266333, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[128], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "attn_fwd", + "cpu_time_us": 0, + "cuda_time_us": 71.546, + "pct_cuda_time": 0.5196675568695539, + "trace": "_attention(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], int32[2], int32[2], 128, 128, True, 0.08838834764831845, None, None, None) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 97.746, + "cuda_time_us": 40.61, + "pct_cuda_time": 0.2949668672528524, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x32x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA1024_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_2_MO40_NTn1_NTA5_NTB0_NTC0_NTD7_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB1_WSGRA0_WSGRB0_WS64_WG32_4_2", + "cpu_time_us": 0, + "cuda_time_us": 40.61, + "pct_cuda_time": 0.2949668672528524, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 4096]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 4096]) <- linear(bfloat16[128, 4096], bfloat16[4096, 4096], None)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 37.816, + "cuda_time_us": 6.569, + "pct_cuda_time": 0.047713305860231175, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.569, + "pct_cuda_time": 0.047713305860231175, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 220.218, + "cuda_time_us": 219.584, + "pct_cuda_time": 1.594927470545441, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 82.533, + "cuda_time_us": 125.857, + "pct_cuda_time": 0.914150332722045, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT96x128x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT6_2_MO40_NTn1_NTA4_NTB0_NTC0_NTD4_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB2_WSGRA0_WSGRB0_WS64_WG16_16_1", + "cpu_time_us": 0, + "cuda_time_us": 125.857, + "pct_cuda_time": 0.914150332722045, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 28672]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 28672]) <- linear(bfloat16[128, 4096], bfloat16[28672, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 25.168, + "cuda_time_us": 12.017, + "pct_cuda_time": 0.08728433498590317, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 12.017, + "pct_cuda_time": 0.08728433498590317, + "trace": "_C::silu_and_mul(bfloat16[128, 14336], bfloat16[128, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 91.456, + "cuda_time_us": 81.71, + "pct_cuda_time": 0.5934928028374925, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT128x64x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB1024_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_4_MO40_NTn1_NTA4_NTB0_NTC0_NTD0_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB4_WSGRA0_WSGRB0_WS64_WG64_4_1", + "cpu_time_us": 0, + "cuda_time_us": 77.264, + "pct_cuda_time": 0.5611997052800884, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + }, + { + "entry": { + "name": "Cijk_SB_BiasS_HAS_ScaleAlphaVec_PostGSU4_VW4", + "cpu_time_us": 0, + "cuda_time_us": 4.446, + "pct_cuda_time": 0.032293097557404137, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 1048.736, + "cuda_time_us": 395.32399999999996, + "pct_cuda_time": 2.87139822284823, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 37.116, + "cuda_time_us": 6.088, + "pct_cuda_time": 0.04421960817127224, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.088, + "pct_cuda_time": 0.04421960817127224, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 724.652, + "cuda_time_us": 153.63, + "pct_cuda_time": 1.115876873086819, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 87.952, + "cuda_time_us": 40.55, + "pct_cuda_time": 0.2945310629673274, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x48x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA2048_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT4_3_MO40_NTn1_NTA0_NTB0_NTC0_NTD0_NTM0_NEPBS8_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW4_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA4_VWB1_WSGRA0_WSGRB0_WS64_WG16_4_4", + "cpu_time_us": 0, + "cuda_time_us": 40.55, + "pct_cuda_time": 0.2945310629673274, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 6144]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 6144]) <- linear(bfloat16[128, 4096], bfloat16[6144, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 71.236, + "cuda_time_us": 20.064, + "pct_cuda_time": 0.1457329530795674, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.245, + "pct_cuda_time": 0.030833153200895313, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False) <- clone(bfloat16[128, 4096], 0) <- contiguous(bfloat16[128, 4096], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.925, + "pct_cuda_time": 0.028508863678095193, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False) <- clone(bfloat16[128, 1024], 0) <- contiguous(bfloat16[128, 1024], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.726, + "pct_cuda_time": 0.034326850889854245, + "trace": "_C::rotary_embedding(int64[128], bfloat16[128, 4096], bfloat16[128, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.764, + "pct_cuda_time": 0.027339455511936388, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.404, + "pct_cuda_time": 0.024724629798786254, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 429.772, + "cuda_time_us": 75.67099999999999, + "pct_cuda_time": 0.549629101499399, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 5.327, + "pct_cuda_time": 0.03869215714986322, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[128], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "attn_fwd", + "cpu_time_us": 0, + "cuda_time_us": 70.344, + "pct_cuda_time": 0.5109369443495358, + "trace": "_attention(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], int32[2], int32[2], 128, 128, True, 0.08838834764831845, None, None, None) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 99.709, + "cuda_time_us": 17.345, + "pct_cuda_time": 0.12598375554052513, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x32x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA1024_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_2_MO40_NTn1_NTA5_NTB0_NTC0_NTD7_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB1_WSGRA0_WSGRB0_WS64_WG32_4_2", + "cpu_time_us": 0, + "cuda_time_us": 17.345, + "pct_cuda_time": 0.12598375554052513, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 4096]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 4096]) <- linear(bfloat16[128, 4096], bfloat16[4096, 4096], None)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 37.706, + "cuda_time_us": 7.049, + "pct_cuda_time": 0.05119974014443135, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 7.049, + "pct_cuda_time": 0.05119974014443135, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 217.555, + "cuda_time_us": 228.557, + "pct_cuda_time": 1.660102001445708, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 83.775, + "cuda_time_us": 135.301, + "pct_cuda_time": 0.9827459272636835, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT96x128x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT6_2_MO40_NTn1_NTA4_NTB0_NTC0_NTD4_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB2_WSGRA0_WSGRB0_WS64_WG16_16_1", + "cpu_time_us": 0, + "cuda_time_us": 135.301, + "pct_cuda_time": 0.9827459272636835, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 28672]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 28672]) <- linear(bfloat16[128, 4096], bfloat16[28672, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 26.739, + "cuda_time_us": 12.177, + "pct_cuda_time": 0.08844647974730323, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 12.177, + "pct_cuda_time": 0.08844647974730323, + "trace": "_C::silu_and_mul(bfloat16[128, 14336], bfloat16[128, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 89.213, + "cuda_time_us": 81.07900000000001, + "pct_cuda_time": 0.5889095944347212, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT128x64x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB1024_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_4_MO40_NTn1_NTA4_NTB0_NTC0_NTD0_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB4_WSGRA0_WSGRB0_WS64_WG64_4_1", + "cpu_time_us": 0, + "cuda_time_us": 76.554, + "pct_cuda_time": 0.5560426879013757, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + }, + { + "entry": { + "name": "Cijk_SB_BiasS_HAS_ScaleAlphaVec_PostGSU4_VW4", + "cpu_time_us": 0, + "cuda_time_us": 4.525, + "pct_cuda_time": 0.032866906533345425, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 1024.96, + "cuda_time_us": 390.858, + "pct_cuda_time": 2.8389598571956514, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 34.632, + "cuda_time_us": 6.169, + "pct_cuda_time": 0.044807943956731024, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.169, + "pct_cuda_time": 0.044807943956731024, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 714.657, + "cuda_time_us": 155.642, + "pct_cuda_time": 1.1304908434614247, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 86.018, + "cuda_time_us": 39.918, + "pct_cuda_time": 0.2899405911597972, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x48x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA2048_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT4_3_MO40_NTn1_NTA0_NTB0_NTC0_NTD0_NTM0_NEPBS8_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW4_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA4_VWB1_WSGRA0_WSGRB0_WS64_WG16_4_4", + "cpu_time_us": 0, + "cuda_time_us": 39.918, + "pct_cuda_time": 0.2899405911597972, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 6144]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 6144]) <- linear(bfloat16[128, 4096], bfloat16[6144, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 72.128, + "cuda_time_us": 20.465, + "pct_cuda_time": 0.1486455783878263, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.206, + "pct_cuda_time": 0.03054988041530405, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False) <- clone(bfloat16[128, 4096], 0) <- contiguous(bfloat16[128, 4096], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.764, + "pct_cuda_time": 0.027339455511936388, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False) <- clone(bfloat16[128, 1024], 0) <- contiguous(bfloat16[128, 1024], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.846, + "pct_cuda_time": 0.03519845946090428, + "trace": "_C::rotary_embedding(int64[128], bfloat16[128, 4096], bfloat16[128, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.205, + "pct_cuda_time": 0.030542617010545302, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.444, + "pct_cuda_time": 0.02501516598913627, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 423.031, + "cuda_time_us": 78.075, + "pct_cuda_time": 0.567090326539435, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 5.287, + "pct_cuda_time": 0.038401620959513195, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[128], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "attn_fwd", + "cpu_time_us": 0, + "cuda_time_us": 72.788, + "pct_cuda_time": 0.5286887055799219, + "trace": "_attention(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], int32[2], int32[2], 128, 128, True, 0.08838834764831845, None, None, None) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 97.416, + "cuda_time_us": 17.184, + "pct_cuda_time": 0.12481434737436634, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x32x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA1024_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_2_MO40_NTn1_NTA5_NTB0_NTC0_NTD7_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB1_WSGRA0_WSGRB0_WS64_WG32_4_2", + "cpu_time_us": 0, + "cuda_time_us": 17.184, + "pct_cuda_time": 0.12481434737436634, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 4096]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 4096]) <- linear(bfloat16[128, 4096], bfloat16[4096, 4096], None)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 37.065, + "cuda_time_us": 6.81, + "pct_cuda_time": 0.04946378640709, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.81, + "pct_cuda_time": 0.04946378640709, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 217.314, + "cuda_time_us": 222.237, + "pct_cuda_time": 1.6141972833704055, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 84.396, + "cuda_time_us": 129.552, + "pct_cuda_time": 0.9409886133056277, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT96x128x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT6_2_MO40_NTn1_NTA4_NTB0_NTC0_NTD4_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB2_WSGRA0_WSGRB0_WS64_WG16_16_1", + "cpu_time_us": 0, + "cuda_time_us": 129.552, + "pct_cuda_time": 0.9409886133056277, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 28672]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 28672]) <- linear(bfloat16[128, 4096], bfloat16[28672, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 26.209, + "cuda_time_us": 11.737, + "pct_cuda_time": 0.08525058165345308, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 11.737, + "pct_cuda_time": 0.08525058165345308, + "trace": "_C::silu_and_mul(bfloat16[128, 14336], bfloat16[128, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 89.914, + "cuda_time_us": 80.94800000000001, + "pct_cuda_time": 0.5879580884113248, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT128x64x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB1024_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_4_MO40_NTn1_NTA4_NTB0_NTC0_NTD0_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB4_WSGRA0_WSGRB0_WS64_WG64_4_1", + "cpu_time_us": 0, + "cuda_time_us": 76.423, + "pct_cuda_time": 0.5550911818779795, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + }, + { + "entry": { + "name": "Cijk_SB_BiasS_HAS_ScaleAlphaVec_PostGSU4_VW4", + "cpu_time_us": 0, + "cuda_time_us": 4.525, + "pct_cuda_time": 0.032866906533345425, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 1027.033, + "cuda_time_us": 392.592, + "pct_cuda_time": 2.8515546010473245, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 33.931, + "cuda_time_us": 6.289, + "pct_cuda_time": 0.04567955252778106, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.289, + "pct_cuda_time": 0.04567955252778106, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 715.388, + "cuda_time_us": 154.98299999999998, + "pct_cuda_time": 1.125704259725408, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 83.955, + "cuda_time_us": 40.339, + "pct_cuda_time": 0.29299848456323113, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x48x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA2048_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT4_3_MO40_NTn1_NTA0_NTB0_NTC0_NTD0_NTM0_NEPBS8_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW4_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA4_VWB1_WSGRA0_WSGRB0_WS64_WG16_4_4", + "cpu_time_us": 0, + "cuda_time_us": 40.339, + "pct_cuda_time": 0.29299848456323113, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 6144]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 6144]) <- linear(bfloat16[128, 4096], bfloat16[6144, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 71.046, + "cuda_time_us": 20.786999999999995, + "pct_cuda_time": 0.15098439472014386, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.486, + "pct_cuda_time": 0.03258363374775415, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False) <- clone(bfloat16[128, 4096], 0) <- contiguous(bfloat16[128, 4096], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.045, + "pct_cuda_time": 0.02938047224914524, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False) <- clone(bfloat16[128, 1024], 0) <- contiguous(bfloat16[128, 1024], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.967, + "pct_cuda_time": 0.03607733143671308, + "trace": "_C::rotary_embedding(int64[128], bfloat16[128, 4096], bfloat16[128, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.005, + "pct_cuda_time": 0.029089936058795226, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.284, + "pct_cuda_time": 0.02385302122773621, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 423.141, + "cuda_time_us": 76.632, + "pct_cuda_time": 0.5566092334725583, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 5.527, + "pct_cuda_time": 0.04014483810161329, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[128], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "attn_fwd", + "cpu_time_us": 0, + "cuda_time_us": 71.105, + "pct_cuda_time": 0.516464395370945, + "trace": "_attention(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], int32[2], int32[2], 128, 128, True, 0.08838834764831845, None, None, None) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 101.501, + "cuda_time_us": 17.225, + "pct_cuda_time": 0.1251121469694751, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x32x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA1024_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_2_MO40_NTn1_NTA5_NTB0_NTC0_NTD7_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB1_WSGRA0_WSGRB0_WS64_WG32_4_2", + "cpu_time_us": 0, + "cuda_time_us": 17.225, + "pct_cuda_time": 0.1251121469694751, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 4096]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 4096]) <- linear(bfloat16[128, 4096], bfloat16[4096, 4096], None)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 36.985, + "cuda_time_us": 6.569, + "pct_cuda_time": 0.047713305860231175, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.569, + "pct_cuda_time": 0.047713305860231175, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 219.528, + "cuda_time_us": 224.75100000000003, + "pct_cuda_time": 1.6324574829339042, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 82.583, + "cuda_time_us": 130.854, + "pct_cuda_time": 0.9504455663015209, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT96x128x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT6_2_MO40_NTn1_NTA4_NTB0_NTC0_NTD4_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB2_WSGRA0_WSGRB0_WS64_WG16_16_1", + "cpu_time_us": 0, + "cuda_time_us": 130.854, + "pct_cuda_time": 0.9504455663015209, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 28672]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 28672]) <- linear(bfloat16[128, 4096], bfloat16[28672, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 25.468, + "cuda_time_us": 12.097, + "pct_cuda_time": 0.0878654073666032, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 12.097, + "pct_cuda_time": 0.0878654073666032, + "trace": "_C::silu_and_mul(bfloat16[128, 14336], bfloat16[128, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 93.249, + "cuda_time_us": 81.8, + "pct_cuda_time": 0.5941465092657802, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT128x64x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB1024_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_4_MO40_NTn1_NTA4_NTB0_NTC0_NTD0_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB4_WSGRA0_WSGRB0_WS64_WG64_4_1", + "cpu_time_us": 0, + "cuda_time_us": 77.234, + "pct_cuda_time": 0.5609818031373259, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + }, + { + "entry": { + "name": "Cijk_SB_BiasS_HAS_ScaleAlphaVec_PostGSU4_VW4", + "cpu_time_us": 0, + "cuda_time_us": 4.566, + "pct_cuda_time": 0.03316470612845418, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 1061.064, + "cuda_time_us": 396.558, + "pct_cuda_time": 2.8803612643205287, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 34.532, + "cuda_time_us": 6.249, + "pct_cuda_time": 0.045389016337431046, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.249, + "pct_cuda_time": 0.045389016337431046, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 720.625, + "cuda_time_us": 155.113, + "pct_cuda_time": 1.126648502344046, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 83.394, + "cuda_time_us": 40.469, + "pct_cuda_time": 0.29394272718186865, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x48x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA2048_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT4_3_MO40_NTn1_NTA0_NTB0_NTC0_NTD0_NTM0_NEPBS8_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW4_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA4_VWB1_WSGRA0_WSGRB0_WS64_WG16_4_4", + "cpu_time_us": 0, + "cuda_time_us": 40.469, + "pct_cuda_time": 0.29394272718186865, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 6144]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 6144]) <- linear(bfloat16[128, 4096], bfloat16[6144, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 72.318, + "cuda_time_us": 20.747, + "pct_cuda_time": 0.1506938585297939, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.286, + "pct_cuda_time": 0.03113095279600408, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False) <- clone(bfloat16[128, 4096], 0) <- contiguous(bfloat16[128, 4096], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.965, + "pct_cuda_time": 0.028799399868445208, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False) <- clone(bfloat16[128, 1024], 0) <- contiguous(bfloat16[128, 1024], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 5.007, + "pct_cuda_time": 0.036367867627063094, + "trace": "_C::rotary_embedding(int64[128], bfloat16[128, 4096], bfloat16[128, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.125, + "pct_cuda_time": 0.029961544629845273, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.364, + "pct_cuda_time": 0.024434093608436236, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 428.259, + "cuda_time_us": 76.432, + "pct_cuda_time": 0.5551565525208082, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 5.447, + "pct_cuda_time": 0.03956376572091325, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[128], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "attn_fwd", + "cpu_time_us": 0, + "cuda_time_us": 70.985, + "pct_cuda_time": 0.515592786799895, + "trace": "_attention(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], int32[2], int32[2], 128, 128, True, 0.08838834764831845, None, None, None) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 101.151, + "cuda_time_us": 17.465, + "pct_cuda_time": 0.1268553641115752, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x32x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA1024_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_2_MO40_NTn1_NTA5_NTB0_NTC0_NTD7_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB1_WSGRA0_WSGRB0_WS64_WG32_4_2", + "cpu_time_us": 0, + "cuda_time_us": 17.465, + "pct_cuda_time": 0.1268553641115752, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 4096]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 4096]) <- linear(bfloat16[128, 4096], bfloat16[4096, 4096], None)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 38.067, + "cuda_time_us": 5.888, + "pct_cuda_time": 0.04276692721952217, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 5.888, + "pct_cuda_time": 0.04276692721952217, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 245.126, + "cuda_time_us": 229.308, + "pct_cuda_time": 1.6655568184195293, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 99.699, + "cuda_time_us": 135.762, + "pct_cuda_time": 0.9860943568574676, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT96x128x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT6_2_MO40_NTn1_NTA4_NTB0_NTC0_NTD4_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB2_WSGRA0_WSGRB0_WS64_WG16_16_1", + "cpu_time_us": 0, + "cuda_time_us": 135.762, + "pct_cuda_time": 0.9860943568574676, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 28672]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 28672]) <- linear(bfloat16[128, 4096], bfloat16[28672, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 29.094, + "cuda_time_us": 13.419, + "pct_cuda_time": 0.0974676284576712, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 13.419, + "pct_cuda_time": 0.0974676284576712, + "trace": "_C::silu_and_mul(bfloat16[128, 14336], bfloat16[128, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 97.496, + "cuda_time_us": 80.127, + "pct_cuda_time": 0.5819948331043908, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT128x64x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB1024_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_4_MO40_NTn1_NTA4_NTB0_NTC0_NTD0_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB4_WSGRA0_WSGRB0_WS64_WG64_4_1", + "cpu_time_us": 0, + "cuda_time_us": 75.722, + "pct_cuda_time": 0.5499995351420953, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + }, + { + "entry": { + "name": "Cijk_SB_BiasS_HAS_ScaleAlphaVec_PostGSU4_VW4", + "cpu_time_us": 0, + "cuda_time_us": 4.405, + "pct_cuda_time": 0.031995297962295374, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 1042.847, + "cuda_time_us": 422.565, + "pct_cuda_time": 3.0692606318813493, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 35.513, + "cuda_time_us": 5.968, + "pct_cuda_time": 0.0433479996002222, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 5.968, + "pct_cuda_time": 0.0433479996002222, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 728.908, + "cuda_time_us": 187.5, + "pct_cuda_time": 1.3618883922656941, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 84.356, + "cuda_time_us": 41.19, + "pct_cuda_time": 0.2991796420129277, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x48x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA2048_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT4_3_MO40_NTn1_NTA0_NTB0_NTC0_NTD0_NTM0_NEPBS8_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW4_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA4_VWB1_WSGRA0_WSGRB0_WS64_WG16_4_4", + "cpu_time_us": 0, + "cuda_time_us": 41.19, + "pct_cuda_time": 0.2991796420129277, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 6144]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 6144]) <- linear(bfloat16[128, 4096], bfloat16[6144, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 72.087, + "cuda_time_us": 21.186, + "pct_cuda_time": 0.1538824932188853, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.326, + "pct_cuda_time": 0.03142148898635409, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False) <- clone(bfloat16[128, 4096], 0) <- contiguous(bfloat16[128, 4096], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.884, + "pct_cuda_time": 0.028211064082986428, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False) <- clone(bfloat16[128, 1024], 0) <- contiguous(bfloat16[128, 1024], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.926, + "pct_cuda_time": 0.03577953184160432, + "trace": "_C::rotary_embedding(int64[128], bfloat16[128, 4096], bfloat16[128, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.486, + "pct_cuda_time": 0.03258363374775415, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.564, + "pct_cuda_time": 0.025886774560186315, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 437.783, + "cuda_time_us": 77.153, + "pct_cuda_time": 0.5603934673518672, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 5.607, + "pct_cuda_time": 0.04072591048231332, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[128], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "attn_fwd", + "cpu_time_us": 0, + "cuda_time_us": 71.546, + "pct_cuda_time": 0.5196675568695539, + "trace": "_attention(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], int32[2], int32[2], 128, 128, True, 0.08838834764831845, None, None, None) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 99.088, + "cuda_time_us": 47.971, + "pct_cuda_time": 0.3484327896820139, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x32x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA1024_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_2_MO40_NTn1_NTA5_NTB0_NTC0_NTD7_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB1_WSGRA0_WSGRB0_WS64_WG32_4_2", + "cpu_time_us": 0, + "cuda_time_us": 47.971, + "pct_cuda_time": 0.3484327896820139, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 4096]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 4096]) <- linear(bfloat16[128, 4096], bfloat16[4096, 4096], None)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 38.818, + "cuda_time_us": 6.729, + "pct_cuda_time": 0.04887545062163123, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.729, + "pct_cuda_time": 0.04887545062163123, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 217.936, + "cuda_time_us": 222.368, + "pct_cuda_time": 1.615148789393802, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 83.975, + "cuda_time_us": 128.06, + "pct_cuda_time": 0.9301516134055723, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT96x128x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT6_2_MO40_NTn1_NTA4_NTB0_NTC0_NTD4_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB2_WSGRA0_WSGRB0_WS64_WG16_16_1", + "cpu_time_us": 0, + "cuda_time_us": 128.06, + "pct_cuda_time": 0.9301516134055723, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 28672]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 28672]) <- linear(bfloat16[128, 4096], bfloat16[28672, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 25.979, + "cuda_time_us": 11.977, + "pct_cuda_time": 0.08699379879555316, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 11.977, + "pct_cuda_time": 0.08699379879555316, + "trace": "_C::silu_and_mul(bfloat16[128, 14336], bfloat16[128, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 88.533, + "cuda_time_us": 82.331, + "pct_cuda_time": 0.5980033771926766, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT128x64x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB1024_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_4_MO40_NTn1_NTA4_NTB0_NTC0_NTD0_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB4_WSGRA0_WSGRB0_WS64_WG64_4_1", + "cpu_time_us": 0, + "cuda_time_us": 77.805, + "pct_cuda_time": 0.5651292072545724, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + }, + { + "entry": { + "name": "Cijk_SB_BiasS_HAS_ScaleAlphaVec_PostGSU4_VW4", + "cpu_time_us": 0, + "cuda_time_us": 4.526, + "pct_cuda_time": 0.032874169938104165, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 1063.577, + "cuda_time_us": 424.59000000000003, + "pct_cuda_time": 3.0839690265178192, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 34.702, + "cuda_time_us": 6.408, + "pct_cuda_time": 0.04654389769407236, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.408, + "pct_cuda_time": 0.04654389769407236, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 747.146, + "cuda_time_us": 182.63500000000002, + "pct_cuda_time": 1.3265519281143736, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 84.056, + "cuda_time_us": 37.716, + "pct_cuda_time": 0.2739465738810289, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x48x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA2048_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT4_3_MO40_NTn1_NTA0_NTB0_NTC0_NTD0_NTM0_NEPBS8_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW4_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA4_VWB1_WSGRA0_WSGRB0_WS64_WG16_4_4", + "cpu_time_us": 0, + "cuda_time_us": 37.716, + "pct_cuda_time": 0.2739465738810289, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 6144]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 6144]) <- linear(bfloat16[128, 4096], bfloat16[6144, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 71.386, + "cuda_time_us": 20.065, + "pct_cuda_time": 0.14574021648432614, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.285, + "pct_cuda_time": 0.031123689391245327, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False) <- clone(bfloat16[128, 4096], 0) <- contiguous(bfloat16[128, 4096], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.885, + "pct_cuda_time": 0.02821832748774518, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False) <- clone(bfloat16[128, 1024], 0) <- contiguous(bfloat16[128, 1024], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.926, + "pct_cuda_time": 0.03577953184160432, + "trace": "_C::rotary_embedding(int64[128], bfloat16[128, 4096], bfloat16[128, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.725, + "pct_cuda_time": 0.02705618272634512, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.244, + "pct_cuda_time": 0.023562485037386196, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 447.207, + "cuda_time_us": 75.632, + "pct_cuda_time": 0.5493458287138079, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 5.488, + "pct_cuda_time": 0.03986156531602202, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[128], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "attn_fwd", + "cpu_time_us": 0, + "cuda_time_us": 70.144, + "pct_cuda_time": 0.5094842633977859, + "trace": "_attention(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], int32[2], int32[2], 128, 128, True, 0.08838834764831845, None, None, None) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 109.013, + "cuda_time_us": 49.222, + "pct_cuda_time": 0.3575193090352106, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x32x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA1024_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_2_MO40_NTn1_NTA5_NTB0_NTC0_NTD7_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB1_WSGRA0_WSGRB0_WS64_WG32_4_2", + "cpu_time_us": 0, + "cuda_time_us": 49.222, + "pct_cuda_time": 0.3575193090352106, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 4096]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 4096]) <- linear(bfloat16[128, 4096], bfloat16[4096, 4096], None)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 39.158, + "cuda_time_us": 6.409, + "pct_cuda_time": 0.04655116109883111, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.409, + "pct_cuda_time": 0.04655116109883111, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 220.139, + "cuda_time_us": 229.138, + "pct_cuda_time": 1.664322039610542, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 85.828, + "cuda_time_us": 134.991, + "pct_cuda_time": 0.980494271788471, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT96x128x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT6_2_MO40_NTn1_NTA4_NTB0_NTC0_NTD4_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB2_WSGRA0_WSGRB0_WS64_WG16_16_1", + "cpu_time_us": 0, + "cuda_time_us": 134.991, + "pct_cuda_time": 0.980494271788471, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 28672]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 28672]) <- linear(bfloat16[128, 4096], bfloat16[28672, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 27.391, + "cuda_time_us": 13.058, + "pct_cuda_time": 0.09484553933976231, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 13.058, + "pct_cuda_time": 0.09484553933976231, + "trace": "_C::silu_and_mul(bfloat16[128, 14336], bfloat16[128, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 89.303, + "cuda_time_us": 81.089, + "pct_cuda_time": 0.5889822284823086, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT128x64x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB1024_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_4_MO40_NTn1_NTA4_NTB0_NTC0_NTD0_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB4_WSGRA0_WSGRB0_WS64_WG64_4_1", + "cpu_time_us": 0, + "cuda_time_us": 76.643, + "pct_cuda_time": 0.5566891309249045, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + }, + { + "entry": { + "name": "Cijk_SB_BiasS_HAS_ScaleAlphaVec_PostGSU4_VW4", + "cpu_time_us": 0, + "cuda_time_us": 4.446, + "pct_cuda_time": 0.032293097557404137, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 1034.955, + "cuda_time_us": 418.89099999999996, + "pct_cuda_time": 3.0425748827977004, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 33.51, + "cuda_time_us": 6.449, + "pct_cuda_time": 0.046841697289181125, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.449, + "pct_cuda_time": 0.046841697289181125, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 718.874, + "cuda_time_us": 184.557, + "pct_cuda_time": 1.3405121920606917, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 82.263, + "cuda_time_us": 40.68, + "pct_cuda_time": 0.29547530558596496, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x48x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA2048_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT4_3_MO40_NTn1_NTA0_NTB0_NTC0_NTD0_NTM0_NEPBS8_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW4_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA4_VWB1_WSGRA0_WSGRB0_WS64_WG16_4_4", + "cpu_time_us": 0, + "cuda_time_us": 40.68, + "pct_cuda_time": 0.29547530558596496, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 6144]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 6144]) <- linear(bfloat16[128, 4096], bfloat16[6144, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 70.856, + "cuda_time_us": 21.307000000000002, + "pct_cuda_time": 0.1547613651946941, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.406, + "pct_cuda_time": 0.03200256136705412, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False) <- clone(bfloat16[128, 4096], 0) <- contiguous(bfloat16[128, 4096], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.925, + "pct_cuda_time": 0.028508863678095193, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False) <- clone(bfloat16[128, 1024], 0) <- contiguous(bfloat16[128, 1024], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.966, + "pct_cuda_time": 0.03607006803195433, + "trace": "_C::rotary_embedding(int64[128], bfloat16[128, 4096], bfloat16[128, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.646, + "pct_cuda_time": 0.03374577850915421, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.364, + "pct_cuda_time": 0.024434093608436236, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 433.757, + "cuda_time_us": 77.15299999999999, + "pct_cuda_time": 0.5603934673518671, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 5.568, + "pct_cuda_time": 0.04044263769672205, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[128], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "attn_fwd", + "cpu_time_us": 0, + "cuda_time_us": 71.585, + "pct_cuda_time": 0.5199508296551452, + "trace": "_attention(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], int32[2], int32[2], 128, 128, True, 0.08838834764831845, None, None, None) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 96.584, + "cuda_time_us": 45.417, + "pct_cuda_time": 0.3298820539281655, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x32x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA1024_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_2_MO40_NTn1_NTA5_NTB0_NTC0_NTD7_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB1_WSGRA0_WSGRB0_WS64_WG32_4_2", + "cpu_time_us": 0, + "cuda_time_us": 45.417, + "pct_cuda_time": 0.3298820539281655, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 4096]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 4096]) <- linear(bfloat16[128, 4096], bfloat16[4096, 4096], None)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 37.055, + "cuda_time_us": 5.968, + "pct_cuda_time": 0.0433479996002222, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 5.968, + "pct_cuda_time": 0.0433479996002222, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 220.73, + "cuda_time_us": 221.917, + "pct_cuda_time": 1.6118729938476055, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 87.371, + "cuda_time_us": 130.864, + "pct_cuda_time": 0.9505182003491082, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT96x128x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT6_2_MO40_NTn1_NTA4_NTB0_NTC0_NTD4_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB2_WSGRA0_WSGRB0_WS64_WG16_16_1", + "cpu_time_us": 0, + "cuda_time_us": 130.864, + "pct_cuda_time": 0.9505182003491082, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 28672]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 28672]) <- linear(bfloat16[128, 4096], bfloat16[28672, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 27.461, + "cuda_time_us": 12.017, + "pct_cuda_time": 0.08728433498590317, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 12.017, + "pct_cuda_time": 0.08728433498590317, + "trace": "_C::silu_and_mul(bfloat16[128, 14336], bfloat16[128, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 87.821, + "cuda_time_us": 79.036, + "pct_cuda_time": 0.5740704585125941, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT128x64x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB1024_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_4_MO40_NTn1_NTA4_NTB0_NTC0_NTD0_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB4_WSGRA0_WSGRB0_WS64_WG64_4_1", + "cpu_time_us": 0, + "cuda_time_us": 74.35, + "pct_cuda_time": 0.5400341438130898, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + }, + { + "entry": { + "name": "Cijk_SB_BiasS_HAS_ScaleAlphaVec_PostGSU4_VW4", + "cpu_time_us": 0, + "cuda_time_us": 4.686, + "pct_cuda_time": 0.03403631469950423, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 1052.211, + "cuda_time_us": 386.531, + "pct_cuda_time": 2.8075311048045384, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 33.61, + "cuda_time_us": 6.248, + "pct_cuda_time": 0.045381752932672305, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.248, + "pct_cuda_time": 0.045381752932672305, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 742.769, + "cuda_time_us": 152.108, + "pct_cuda_time": 1.104821971044001, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 83.054, + "cuda_time_us": 39.668, + "pct_cuda_time": 0.2881247399701096, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x48x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA2048_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT4_3_MO40_NTn1_NTA0_NTB0_NTC0_NTD0_NTM0_NEPBS8_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW4_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA4_VWB1_WSGRA0_WSGRB0_WS64_WG16_4_4", + "cpu_time_us": 0, + "cuda_time_us": 39.668, + "pct_cuda_time": 0.2881247399701096, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 6144]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 6144]) <- linear(bfloat16[128, 4096], bfloat16[6144, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 70.245, + "cuda_time_us": 20.065, + "pct_cuda_time": 0.14574021648432614, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 4.205, + "pct_cuda_time": 0.030542617010545302, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False) <- clone(bfloat16[128, 4096], 0) <- contiguous(bfloat16[128, 4096], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.765, + "pct_cuda_time": 0.027346718916695135, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False) <- clone(bfloat16[128, 1024], 0) <- contiguous(bfloat16[128, 1024], 0)" + }, + "children": [] + }, + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.846, + "pct_cuda_time": 0.03519845946090428, + "trace": "_C::rotary_embedding(int64[128], bfloat16[128, 4096], bfloat16[128, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.805, + "pct_cuda_time": 0.027637255107045153, + "trace": "copy_(bfloat16[128, 4096], bfloat16[128, 4096], False)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl_nocast(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#12}::operator()() const::{lambda(c10::BFloat16)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.444, + "pct_cuda_time": 0.02501516598913627, + "trace": "copy_(bfloat16[128, 1024], bfloat16[128, 1024], False)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 435.951, + "cuda_time_us": 74.46900000000001, + "pct_cuda_time": 0.5408984889793813, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 5.287, + "pct_cuda_time": 0.038401620959513195, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[128], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "attn_fwd", + "cpu_time_us": 0, + "cuda_time_us": 69.182, + "pct_cuda_time": 0.502496868019868, + "trace": "_attention(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], int32[2], int32[2], 128, 128, True, 0.08838834764831845, None, None, None) <- vllm::unified_attention_with_output(bfloat16[128, 32, 128], bfloat16[128, 8, 128], bfloat16[128, 8, 128], bfloat16[128, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 117.095, + "cuda_time_us": 17.906, + "pct_cuda_time": 0.13005852561018408, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT64x32x256_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA1024_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_2_MO40_NTn1_NTA5_NTB0_NTC0_NTD7_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB1_WSGRA0_WSGRB0_WS64_WG32_4_2", + "cpu_time_us": 0, + "cuda_time_us": 17.906, + "pct_cuda_time": 0.13005852561018408, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 4096]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 4096]) <- linear(bfloat16[128, 4096], bfloat16[4096, 4096], None)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 38.788, + "cuda_time_us": 6.168, + "pct_cuda_time": 0.044800680551972276, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.168, + "pct_cuda_time": 0.044800680551972276, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 215.321, + "cuda_time_us": 222.007, + "pct_cuda_time": 1.6125267002758932, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 83.905, + "cuda_time_us": 128.16, + "pct_cuda_time": 0.9308779538814472, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT96x128x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB512_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT6_2_MO40_NTn1_NTA4_NTB0_NTC0_NTD4_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB2_WSGRA0_WSGRB0_WS64_WG16_16_1", + "cpu_time_us": 0, + "cuda_time_us": 128.16, + "pct_cuda_time": 0.9308779538814472, + "trace": "mm(bfloat16[128, 4096], bfloat16[4096, 28672]) <- matmul(bfloat16[128, 4096], bfloat16[4096, 28672]) <- linear(bfloat16[128, 4096], bfloat16[28672, 4096], None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 25.588, + "cuda_time_us": 11.897, + "pct_cuda_time": 0.08641272641485313, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 11.897, + "pct_cuda_time": 0.08641272641485313, + "trace": "_C::silu_and_mul(bfloat16[128, 14336], bfloat16[128, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 88.012, + "cuda_time_us": 81.95, + "pct_cuda_time": 0.5952360199795927, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Cijk_Alik_Bljk_BBS_BH_Bias_HAS_SAV_UserArgs_MT128x64x128_MI16x16x1_SN_LDSB1_AFC1_AFEM1_AFEM1_ASEM1_CLR1_CADS0_DTVA0_DTVB0_EPS0_FDSI0_GRPM1_GRVWA8_GRVWB8_GSUAMB_GLS0_ISA942_IU1_K1_LBSPPA512_LBSPPB1024_LBSPPM0_LPA16_LPB16_LPM0_LRVW8_LWPMn1_MIAV0_MIWT2_4_MO40_NTn1_NTA4_NTB0_NTC0_NTD0_NTM0_NEPBS16_NLCA1_NLCB1_ONLL1_PGR2_PLR1_PKA1_SIA3_SS1_SPO0_SRVW0_SSO0_SVW2_SK0_SKXCCM0_TLDS1_ULSGRO0_USL1_UIOFGRO0_USFGROn1_VSn1_VWA2_VWB4_WSGRA0_WSGRB0_WS64_WG64_4_1", + "cpu_time_us": 0, + "cuda_time_us": 77.465, + "pct_cuda_time": 0.5626596496365973, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + }, + { + "entry": { + "name": "Cijk_SB_BiasS_HAS_ScaleAlphaVec_PostGSU4_VW4", + "cpu_time_us": 0, + "cuda_time_us": 4.485, + "pct_cuda_time": 0.0325763703429954, + "trace": "mm(bfloat16[128, 14336], bfloat16[14336, 4096]) <- matmul(bfloat16[128, 14336], bfloat16[14336, 4096]) <- linear(bfloat16[128, 14336], bfloat16[4096, 14336], None)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 35.944, + "cuda_time_us": 6.169, + "pct_cuda_time": 0.044807943956731024, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 6.169, + "pct_cuda_time": 0.044807943956731024, + "trace": "_C::fused_add_rms_norm(bfloat16[128, 4096], bfloat16[128, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "LogitsProcessor", + "cpu_time_us": 146.098, + "cuda_time_us": 262.761, + "pct_cuda_time": 1.9085394978140058, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void at::native::(anonymous namespace)::indexSelectSmallIndex(at::cuda::detail::TensorInfo, at::cuda::detail::TensorInfo, at::cuda::detail::TensorInfo, int, int, unsigned int, long)", + "cpu_time_us": 0, + "cuda_time_us": 3.004, + "pct_cuda_time": 0.021819267895286105, + "trace": "index_select(bfloat16[128, 4096], 0, int64[1])" + }, + "children": [] + }, + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 259.757, + "pct_cuda_time": 1.8867202299187196, + "trace": "_rocm_C::wvSplitK(bfloat16[128256, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Sampler", + "cpu_time_us": 786.965, + "cuda_time_us": 209.138, + "pct_cuda_time": 1.5190539444355344, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Memcpy DtoD (Device -> Device)", + "cpu_time_us": 0, + "cuda_time_us": 4.646, + "pct_cuda_time": 0.03374577850915421, + "trace": "copy_(bfloat16[1], bfloat16[1], True) <- _to_copy(bfloat16[1], 15, 0, None, None, True, None) <- to(bfloat16[1], 15, 0, None, None, True, False, None)" + }, + "children": [] + }, + { + "entry": { + "name": "Memcpy DtoD (Device -> Device)", + "cpu_time_us": 0, + "cuda_time_us": 4.646, + "pct_cuda_time": 0.03374577850915421, + "trace": "copy_(bfloat16[1], bfloat16[1], True) <- _to_copy(bfloat16[1], 15, 0, None, None, True, None) <- to(bfloat16[1], 15, 0, None, None, True, False, None)" + }, + "children": [] + }, + { + "entry": { + "name": "Memcpy DtoD (Device -> Device)", + "cpu_time_us": 0, + "cuda_time_us": 4.606, + "pct_cuda_time": 0.033455242318804194, + "trace": "copy_(int32[1], int32[1], True) <- _to_copy(int32[1], 3, 0, None, None, True, None) <- to(int32[1], 3, 0, None, None, True, False, None)" + }, + "children": [] + }, + { + "entry": { + "name": "Memcpy DtoD (Device -> Device)", + "cpu_time_us": 0, + "cuda_time_us": 4.566, + "pct_cuda_time": 0.03316470612845418, + "trace": "copy_(bfloat16[1], bfloat16[1], True) <- _to_copy(bfloat16[1], 15, 0, None, None, True, None) <- to(bfloat16[1], 15, 0, None, None, True, False, None)" + }, + "children": [] + }, + { + "entry": { + "name": "Memcpy DtoD (Device -> Device)", + "cpu_time_us": 0, + "cuda_time_us": 4.525, + "pct_cuda_time": 0.032866906533345425, + "trace": "copy_(bfloat16[1], bfloat16[1], True) <- _to_copy(bfloat16[1], 15, 0, None, None, True, None) <- to(bfloat16[1], 15, 0, None, None, True, False, None)" + }, + "children": [] + }, + { + "entry": { + "name": "Memcpy DtoD (Device -> Device)", + "cpu_time_us": 0, + "cuda_time_us": 4.526, + "pct_cuda_time": 0.032874169938104165, + "trace": "copy_(bfloat16[1], bfloat16[1], True) <- _to_copy(bfloat16[1], 15, 0, None, None, True, None) <- to(bfloat16[1], 15, 0, None, None, True, False, None)" + }, + "children": [] + }, + { + "entry": { + "name": "Memcpy DtoD (Device -> Device)", + "cpu_time_us": 0, + "cuda_time_us": 3.965, + "pct_cuda_time": 0.028799399868445208, + "trace": "copy_(bfloat16[1], bfloat16[1], True) <- _to_copy(bfloat16[1], 15, 0, None, None, True, None) <- to(bfloat16[1], 15, 0, None, None, True, False, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<512, 1, at::native::gpu_kernel_impl(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#7}::operator()() const::{lambda(float)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#7}::operator()() const::{lambda(float)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 3.083, + "pct_cuda_time": 0.022393076871227387, + "trace": "copy_(float32[1, 128256], bfloat16[1, 128256], False) <- _to_copy(bfloat16[1, 128256], 6, None, None, None, False, None) <- to(bfloat16[1, 128256], 6, False, False, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl > >(at::TensorIteratorBase&, at::native::BinaryFunctor > const&)::{lambda(int)#2}>(int, at::native::gpu_kernel_impl > >(at::TensorIteratorBase&, at::native::BinaryFunctor > const&)::{lambda(int)#2})", + "cpu_time_us": 0, + "cuda_time_us": 13.54, + "pct_cuda_time": 0.09834650043347999, + "trace": "div_(float32[1, 128256], bfloat16[1, 1])" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::(anonymous namespace)::cunn_SoftMaxForward<4, float, float, float, at::native::(anonymous namespace)::SoftMaxForwardEpilogue>(float*, float const*, int)", + "cpu_time_us": 0, + "cuda_time_us": 64.616, + "pct_cuda_time": 0.4693321618914138, + "trace": "_softmax(float32[1, 128256], -1, False) <- softmax(float32[1, 128256], -1, 6)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::(anonymous namespace)::cunn_SoftMaxForward<4, float, float, float, at::native::(anonymous namespace)::LogSoftMaxForwardEpilogue>(float*, float const*, int)", + "cpu_time_us": 0, + "cuda_time_us": 35.932, + "pct_cuda_time": 0.2609886597914183, + "trace": "_log_softmax(float32[1, 128256], -1, False) <- log_softmax(float32[1, 128256], -1, 6)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<512, 1, at::native::gpu_kernel_impl(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#4}::operator()() const::{lambda(long)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#4}::operator()() const::{lambda(long)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 5.167, + "pct_cuda_time": 0.03753001238846315, + "trace": "copy_(int64[1], int32[1], False) <- _to_copy(int32[1], 4, None, None, None, False, None) <- to(int32[1], 4, False, False, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::index_elementwise_kernel<128, 4, at::native::gpu_index_kernel >(at::TensorIteratorBase&, c10::ArrayRef, c10::ArrayRef)::{lambda(char*, char const*, long)#1}>(at::TensorIteratorBase&, c10::ArrayRef, c10::ArrayRef, at::native::index_kernel_impl >(at::TensorIteratorBase&, c10::ArrayRef, c10::ArrayRef)::{lambda(char*, char const*, long)#1} const&)::{lambda(int)#1}>(long, at::native::gpu_index_kernel >(at::TensorIteratorBase&, c10::ArrayRef, c10::ArrayRef)::{lambda(char*, char const*, long)#1}>(at::TensorIteratorBase&, c10::ArrayRef, c10::ArrayRef, at::native::index_kernel_impl >(at::TensorIteratorBase&, c10::ArrayRef, c10::ArrayRef)::{lambda(char*, char const*, long)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 6.649, + "pct_cuda_time": 0.0482943782409312, + "trace": "index(float32[1, 128256], None)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::reduce_kernel<512, 1, at::native::ReduceOp, unsigned int, long, 4, 4> >(at::native::ReduceOp, unsigned int, long, 4, 4>)", + "cpu_time_us": 0, + "cuda_time_us": 44.025, + "pct_cuda_time": 0.31977139450398495, + "trace": "argmax(float32[1, 128256], -1, False)" + }, + "children": [] + }, + { + "entry": { + "name": "Memcpy DtoH (Device -> Host)", + "cpu_time_us": 0, + "cuda_time_us": 4.646, + "pct_cuda_time": 0.03374577850915421, + "trace": "copy_(int64[1], int64[1], False) <- _to_copy(int64[1], 4, 0, None, None, False, None) <- to(int64[1], 4, 0, None, None, False, False, None)" + }, + "children": [] + } + ] + } + ] + }, + "decode_1": { + "metadata": { + "num_running_seqs": 1 + }, + "summary_stats": [ + { + "entry": { + "name": "LlamaForCausalLM", + "cuda_time_us": 5324.1269999999995, + "pct_cuda_time": 92.26684028963194, + "invocations": 1 + }, + "children": [ + { + "entry": { + "name": "VocabParallelEmbedding(weight=bfloat16[128256, 4096])", + "cuda_time_us": 3.604, + "pct_cuda_time": 0.062457130042884695, + "invocations": 1 + }, + "children": [ + { + "entry": { + "name": "void at::native::(anonymous namespace)::indexSelectSmallIndex(at::cuda::detail::TensorInfo, at::cuda::detail::TensorInfo, at::cuda::detail::TensorInfo, int, int, unsigned int, long)", + "cuda_time_us": 3.604, + "pct_cuda_time": 0.062457130042884695, + "invocations": 1 + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cuda_time_us": 5316.077, + "pct_cuda_time": 92.12733421392574, + "invocations": 32 + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cuda_time_us": 300.30399999999986, + "pct_cuda_time": 5.20425249178647, + "invocations": 64 + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::rms_norm_kernel(c10::BFloat16*, c10::BFloat16 const*, c10::BFloat16 const*, float, int, unsigned long, unsigned long)", + "cuda_time_us": 5.207, + "pct_cuda_time": 0.09023703555308007, + "invocations": 1 + }, + "children": [] + }, + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cuda_time_us": 295.0969999999999, + "pct_cuda_time": 5.114015456233391, + "invocations": 63 + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cuda_time_us": 1570.4170000000004, + "pct_cuda_time": 27.215243837557402, + "invocations": 32 + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cuda_time_us": 546.232, + "pct_cuda_time": 9.46617176958518, + "invocations": 32 + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cuda_time_us": 546.232, + "pct_cuda_time": 9.46617176958518, + "invocations": 32 + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cuda_time_us": 123.674, + "pct_cuda_time": 2.1432639014771704, + "invocations": 32 + }, + "children": [ + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cuda_time_us": 123.674, + "pct_cuda_time": 2.1432639014771704, + "invocations": 32 + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cuda_time_us": 439.88599999999997, + "pct_cuda_time": 7.623201194795887, + "invocations": 32 + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cuda_time_us": 93.51199999999997, + "pct_cuda_time": 1.6205580312348034, + "invocations": 32 + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_QKV_mfma4_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0, __hip_bfloat16, 16, 128, 256, false, 4>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16 const*, int, float, int const*, int const*, int const*, int, float const*, int, int, int, float*, float*, __hip_bfloat16*, __hip_bfloat16*, int, float const*, float const*, float const*)", + "cuda_time_us": 232.39399999999995, + "pct_cuda_time": 4.027375771139329, + "invocations": 32 + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_reduce_kernel<__hip_bfloat16, __hip_bfloat16, 128, 128, 256, 1>(__hip_bfloat16*, float const*, float const*, __hip_bfloat16 const*, int const*, int const*, int, float const*)", + "cuda_time_us": 113.97999999999999, + "pct_cuda_time": 1.9752673924217528, + "invocations": 32 + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cuda_time_us": 460.62500000000006, + "pct_cuda_time": 7.982606971699157, + "invocations": 32 + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cuda_time_us": 460.62500000000006, + "pct_cuda_time": 7.982606971699157, + "invocations": 32 + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cuda_time_us": 3445.3559999999993, + "pct_cuda_time": 59.70783788458185, + "invocations": 32 + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cuda_time_us": 2031.482, + "pct_cuda_time": 35.20547598606534, + "invocations": 32 + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cuda_time_us": 2031.482, + "pct_cuda_time": 35.20547598606534, + "invocations": 32 + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cuda_time_us": 262.19899999999996, + "pct_cuda_time": 4.543894850198202, + "invocations": 32 + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cuda_time_us": 262.19899999999996, + "pct_cuda_time": 4.543894850198202, + "invocations": 32 + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cuda_time_us": 1151.675, + "pct_cuda_time": 19.958467048318322, + "invocations": 32 + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cuda_time_us": 1151.675, + "pct_cuda_time": 19.958467048318322, + "invocations": 32 + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cuda_time_us": 4.446, + "pct_cuda_time": 0.07704894566333666, + "invocations": 1 + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cuda_time_us": 4.446, + "pct_cuda_time": 0.07704894566333666, + "invocations": 1 + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "LogitsProcessor", + "cuda_time_us": 264.814, + "pct_cuda_time": 4.58921266236861, + "invocations": 1 + }, + "children": [ + { + "entry": { + "name": "void at::native::(anonymous namespace)::indexSelectSmallIndex(at::cuda::detail::TensorInfo, at::cuda::detail::TensorInfo, at::cuda::detail::TensorInfo, int, int, unsigned int, long)", + "cuda_time_us": 2.603, + "pct_cuda_time": 0.045109852802893695, + "invocations": 1 + }, + "children": [] + }, + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cuda_time_us": 262.211, + "pct_cuda_time": 4.544102809565715, + "invocations": 1 + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Sampler", + "cuda_time_us": 181.41700000000003, + "pct_cuda_time": 3.143947047999449, + "invocations": 1 + }, + "children": [ + { + "entry": { + "name": "Memcpy DtoD (Device -> Device)", + "cuda_time_us": 16.618000000000002, + "pct_cuda_time": 0.2879890641100605, + "invocations": 7 + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<512, 1, at::native::gpu_kernel_impl(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#7}::operator()() const::{lambda(float)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#7}::operator()() const::{lambda(float)#1} const&)::{lambda(int)#1})", + "cuda_time_us": 2.883, + "pct_cuda_time": 0.04996223804484922, + "invocations": 1 + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl > >(at::TensorIteratorBase&, at::native::BinaryFunctor > const&)::{lambda(int)#2}>(int, at::native::gpu_kernel_impl > >(at::TensorIteratorBase&, at::native::BinaryFunctor > const&)::{lambda(int)#2})", + "cuda_time_us": 8.612, + "pct_cuda_time": 0.14924550608471782, + "invocations": 1 + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::(anonymous namespace)::cunn_SoftMaxForward<4, float, float, float, at::native::(anonymous namespace)::SoftMaxForwardEpilogue>(float*, float const*, int)", + "cuda_time_us": 64.015, + "pct_cuda_time": 1.1093765759420822, + "invocations": 1 + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::(anonymous namespace)::cunn_SoftMaxForward<4, float, float, float, at::native::(anonymous namespace)::LogSoftMaxForwardEpilogue>(float*, float const*, int)", + "cuda_time_us": 35.612, + "pct_cuda_time": 0.6171540829875721, + "invocations": 1 + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<512, 1, at::native::gpu_kernel_impl(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#4}::operator()() const::{lambda(long)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#4}::operator()() const::{lambda(long)#1} const&)::{lambda(int)#1})", + "cuda_time_us": 2.323, + "pct_cuda_time": 0.04025746756093816, + "invocations": 1 + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::index_elementwise_kernel<128, 4, at::native::gpu_index_kernel >(at::TensorIteratorBase&, c10::ArrayRef, c10::ArrayRef)::{lambda(char*, char const*, long)#1}>(at::TensorIteratorBase&, c10::ArrayRef, c10::ArrayRef, at::native::index_kernel_impl >(at::TensorIteratorBase&, c10::ArrayRef, c10::ArrayRef)::{lambda(char*, char const*, long)#1} const&)::{lambda(int)#1}>(long, at::native::gpu_index_kernel >(at::TensorIteratorBase&, c10::ArrayRef, c10::ArrayRef)::{lambda(char*, char const*, long)#1}>(at::TensorIteratorBase&, c10::ArrayRef, c10::ArrayRef, at::native::index_kernel_impl >(at::TensorIteratorBase&, c10::ArrayRef, c10::ArrayRef)::{lambda(char*, char const*, long)#1} const&)::{lambda(int)#1})", + "cuda_time_us": 5.287, + "pct_cuda_time": 0.09162343133649595, + "invocations": 1 + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::reduce_kernel<512, 1, at::native::ReduceOp, unsigned int, long, 4, 4> >(at::native::ReduceOp, unsigned int, long, 4, 4>)", + "cuda_time_us": 41.341, + "pct_cuda_time": 0.7164373510274407, + "invocations": 1 + }, + "children": [] + }, + { + "entry": { + "name": "Memcpy DtoH (Device -> Host)", + "cuda_time_us": 4.726, + "pct_cuda_time": 0.0819013309052922, + "invocations": 1 + }, + "children": [] + } + ] + } + ], + "model_stats": [ + { + "entry": { + "name": "LlamaForCausalLM", + "cpu_time_us": 20777.759, + "cuda_time_us": 5324.1269999999995, + "pct_cuda_time": 92.26684028963194, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "VocabParallelEmbedding(weight=bfloat16[128256, 4096])", + "cpu_time_us": 193.799, + "cuda_time_us": 3.604, + "pct_cuda_time": 0.062457130042884695, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void at::native::(anonymous namespace)::indexSelectSmallIndex(at::cuda::detail::TensorInfo, at::cuda::detail::TensorInfo, at::cuda::detail::TensorInfo, int, int, unsigned int, long)", + "cpu_time_us": 0, + "cuda_time_us": 3.604, + "pct_cuda_time": 0.062457130042884695, + "trace": "index_select(bfloat16[128256, 4096], 0, int64[1]) <- embedding(bfloat16[128256, 4096], int64[1], -1, False, False)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 1328.624, + "cuda_time_us": 186.187, + "pct_cuda_time": 3.2266108965856195, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 171.496, + "cuda_time_us": 5.207, + "pct_cuda_time": 0.09023703555308007, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::rms_norm_kernel(c10::BFloat16*, c10::BFloat16 const*, c10::BFloat16 const*, float, int, unsigned long, unsigned long)", + "cpu_time_us": 0, + "cuda_time_us": 5.207, + "pct_cuda_time": 0.09023703555308007, + "trace": "_C::rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 837.029, + "cuda_time_us": 57.039, + "pct_cuda_time": 0.9884828636282187, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 213.539, + "cuda_time_us": 19.388, + "pct_cuda_time": 0.3359930181108348, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 19.388, + "pct_cuda_time": 0.3359930181108348, + "trace": "_rocm_C::wvSplitK(bfloat16[6144, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 84.326, + "cuda_time_us": 4.807, + "pct_cuda_time": 0.08330505663600077, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.807, + "pct_cuda_time": 0.08330505663600077, + "trace": "_C::rotary_embedding(int64[1], bfloat16[1, 4096], bfloat16[1, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 345.406, + "cuda_time_us": 17.663, + "pct_cuda_time": 0.3060988590309302, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 3.845, + "pct_cuda_time": 0.066633647340425, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[1], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_QKV_mfma4_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0, __hip_bfloat16, 16, 128, 256, false, 4>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16 const*, int, float, int const*, int const*, int const*, int, float const*, int, int, int, float*, float*, __hip_bfloat16*, __hip_bfloat16*, int, float const*, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 9.933, + "pct_cuda_time": 0.17213836645837227, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_reduce_kernel<__hip_bfloat16, __hip_bfloat16, 128, 128, 256, 1>(__hip_bfloat16*, float const*, float const*, __hip_bfloat16 const*, int const*, int const*, int, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 3.885, + "pct_cuda_time": 0.06732684523213292, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 80.31, + "cuda_time_us": 15.181, + "pct_cuda_time": 0.263085929850453, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 15.181, + "pct_cuda_time": 0.263085929850453, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 69.914, + "cuda_time_us": 5.006, + "pct_cuda_time": 0.08675371614724772, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 5.006, + "pct_cuda_time": 0.08675371614724772, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 209.032, + "cuda_time_us": 118.935, + "pct_cuda_time": 2.061137281257073, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 75.853, + "cuda_time_us": 71.626, + "pct_cuda_time": 1.241274804786809, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 71.626, + "pct_cuda_time": 1.241274804786809, + "trace": "_rocm_C::wvSplitK(bfloat16[28672, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 41.321, + "cuda_time_us": 9.213, + "pct_cuda_time": 0.15966080440762948, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 9.213, + "pct_cuda_time": 0.15966080440762948, + "trace": "_C::silu_and_mul(bfloat16[1, 14336], bfloat16[1, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 61.853, + "cuda_time_us": 38.096, + "pct_cuda_time": 0.6602016720626347, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 38.096, + "pct_cuda_time": 0.6602016720626347, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 14336], bfloat16[1, 14336], 304)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 666.555, + "cuda_time_us": 182.502, + "pct_cuda_time": 3.162750040812026, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 31.857, + "cuda_time_us": 5.007, + "pct_cuda_time": 0.08677104609454042, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 5.007, + "pct_cuda_time": 0.08677104609454042, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 401.149, + "cuda_time_us": 52.952999999999996, + "pct_cuda_time": 0.9176726989902533, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 65.357, + "cuda_time_us": 19.027, + "pct_cuda_time": 0.3297369071381707, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 19.027, + "pct_cuda_time": 0.3297369071381707, + "trace": "_rocm_C::wvSplitK(bfloat16[6144, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 29.634, + "cuda_time_us": 3.885, + "pct_cuda_time": 0.06732684523213292, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 3.885, + "pct_cuda_time": 0.06732684523213292, + "trace": "_C::rotary_embedding(int64[1], bfloat16[1, 4096], bfloat16[1, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 193.589, + "cuda_time_us": 15.26, + "pct_cuda_time": 0.26445499568657616, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 3.525, + "pct_cuda_time": 0.06108806420676153, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[1], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_QKV_mfma4_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0, __hip_bfloat16, 16, 128, 256, false, 4>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16 const*, int, float, int const*, int const*, int const*, int, float const*, int, int, int, float*, float*, __hip_bfloat16*, __hip_bfloat16*, int, float const*, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 7.53, + "pct_cuda_time": 0.13049450311401825, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_reduce_kernel<__hip_bfloat16, __hip_bfloat16, 128, 128, 256, 1>(__hip_bfloat16*, float const*, float const*, __hip_bfloat16 const*, int const*, int const*, int, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 4.205, + "pct_cuda_time": 0.07287242836579638, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 70.876, + "cuda_time_us": 14.781, + "pct_cuda_time": 0.2561539509333737, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 14.781, + "pct_cuda_time": 0.2561539509333737, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 37.957, + "cuda_time_us": 4.926, + "pct_cuda_time": 0.08536732036383185, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.926, + "pct_cuda_time": 0.08536732036383185, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 167.33, + "cuda_time_us": 119.616, + "pct_cuda_time": 2.072938975363401, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 58.567, + "cuda_time_us": 73.429, + "pct_cuda_time": 1.272520699755544, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 73.429, + "pct_cuda_time": 1.272520699755544, + "trace": "_rocm_C::wvSplitK(bfloat16[28672, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 22.794, + "cuda_time_us": 8.491, + "pct_cuda_time": 0.1471485824623013, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 8.491, + "pct_cuda_time": 0.1471485824623013, + "trace": "_C::silu_and_mul(bfloat16[1, 14336], bfloat16[1, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 67.22, + "cuda_time_us": 37.696, + "pct_cuda_time": 0.6532696931455554, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 37.696, + "pct_cuda_time": 0.6532696931455554, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 14336], bfloat16[1, 14336], 304)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 647.878, + "cuda_time_us": 195.239, + "pct_cuda_time": 3.3834815794791244, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 32.238, + "cuda_time_us": 5.287, + "pct_cuda_time": 0.09162343133649595, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 5.287, + "pct_cuda_time": 0.09162343133649595, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 399.276, + "cuda_time_us": 81.435, + "pct_cuda_time": 1.4112642577808867, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 57.456, + "cuda_time_us": 18.908, + "pct_cuda_time": 0.3276746434103396, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 18.908, + "pct_cuda_time": 0.3276746434103396, + "trace": "_rocm_C::wvSplitK(bfloat16[6144, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 26.149, + "cuda_time_us": 3.804, + "pct_cuda_time": 0.06592311950142435, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 3.804, + "pct_cuda_time": 0.06592311950142435, + "trace": "_C::rotary_embedding(int64[1], bfloat16[1, 4096], bfloat16[1, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 214.991, + "cuda_time_us": 45.625, + "pct_cuda_time": 0.7906788452293602, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 3.044, + "pct_cuda_time": 0.05275235955897364, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[1], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_QKV_mfma4_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0, __hip_bfloat16, 16, 128, 256, false, 4>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16 const*, int, float, int const*, int const*, int const*, int, float const*, int, int, int, float*, float*, __hip_bfloat16*, __hip_bfloat16*, int, float const*, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 38.296, + "pct_cuda_time": 0.6636676615211743, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_reduce_kernel<__hip_bfloat16, __hip_bfloat16, 128, 128, 256, 1>(__hip_bfloat16*, float const*, float const*, __hip_bfloat16 const*, int const*, int const*, int, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 4.285, + "pct_cuda_time": 0.07425882414921225, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 63.064, + "cuda_time_us": 13.098, + "pct_cuda_time": 0.2269876496397624, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 13.098, + "pct_cuda_time": 0.2269876496397624, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 31.587, + "cuda_time_us": 4.846, + "pct_cuda_time": 0.08398092458041599, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.846, + "pct_cuda_time": 0.08398092458041599, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 156.904, + "cuda_time_us": 103.67099999999999, + "pct_cuda_time": 1.796612965781326, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 56.654, + "cuda_time_us": 62.973, + "pct_cuda_time": 1.0913187708630903, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 62.973, + "pct_cuda_time": 1.0913187708630903, + "trace": "_rocm_C::wvSplitK(bfloat16[28672, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 21.232, + "cuda_time_us": 8.812, + "pct_cuda_time": 0.15271149554325747, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 8.812, + "pct_cuda_time": 0.15271149554325747, + "trace": "_C::silu_and_mul(bfloat16[1, 14336], bfloat16[1, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 61.943, + "cuda_time_us": 31.886, + "pct_cuda_time": 0.5525826993749782, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 31.886, + "pct_cuda_time": 0.5525826993749782, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 14336], bfloat16[1, 14336], 304)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 583.261, + "cuda_time_us": 195.47, + "pct_cuda_time": 3.387484797303738, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 27.341, + "cuda_time_us": 4.726, + "pct_cuda_time": 0.0819013309052922, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.726, + "pct_cuda_time": 0.0819013309052922, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 350.854, + "cuda_time_us": 47.184000000000005, + "pct_cuda_time": 0.8176962330586769, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 52.448, + "cuda_time_us": 16.744, + "pct_cuda_time": 0.29017263746894045, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 16.744, + "pct_cuda_time": 0.29017263746894045, + "trace": "_rocm_C::wvSplitK(bfloat16[6144, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 24.507, + "cuda_time_us": 4.125, + "pct_cuda_time": 0.07148603258238051, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.125, + "pct_cuda_time": 0.07148603258238051, + "trace": "_C::rotary_embedding(int64[1], bfloat16[1, 4096], bfloat16[1, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 168.762, + "cuda_time_us": 13.858, + "pct_cuda_time": 0.24015840958221316, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 3.204, + "pct_cuda_time": 0.055525151125805375, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[1], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_QKV_mfma4_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0, __hip_bfloat16, 16, 128, 256, false, 4>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16 const*, int, float, int const*, int const*, int const*, int, float const*, int, int, int, float*, float*, __hip_bfloat16*, __hip_bfloat16*, int, float const*, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 6.649, + "pct_cuda_time": 0.11522681954915105, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_reduce_kernel<__hip_bfloat16, __hip_bfloat16, 128, 128, 256, 1>(__hip_bfloat16*, float const*, float const*, __hip_bfloat16 const*, int const*, int const*, int, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 4.005, + "pct_cuda_time": 0.06940643890725672, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 70.606, + "cuda_time_us": 12.457, + "pct_cuda_time": 0.2158791534251428, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 12.457, + "pct_cuda_time": 0.2158791534251428, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 28.693, + "cuda_time_us": 4.846, + "pct_cuda_time": 0.08398092458041599, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.846, + "pct_cuda_time": 0.08398092458041599, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 153.209, + "cuda_time_us": 138.714, + "pct_cuda_time": 2.403906308759353, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 57.787, + "cuda_time_us": 61.781, + "pct_cuda_time": 1.070661473690194, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 61.781, + "pct_cuda_time": 1.070661473690194, + "trace": "_rocm_C::wvSplitK(bfloat16[28672, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 19.739, + "cuda_time_us": 8.572, + "pct_cuda_time": 0.14855230819300985, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 8.572, + "pct_cuda_time": 0.14855230819300985, + "trace": "_C::silu_and_mul(bfloat16[1, 14336], bfloat16[1, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 54.762, + "cuda_time_us": 68.361, + "pct_cuda_time": 1.184692526876149, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 68.361, + "pct_cuda_time": 1.184692526876149, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 14336], bfloat16[1, 14336], 304)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 603.3, + "cuda_time_us": 165.889, + "pct_cuda_time": 2.8748476264384295, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 34.602, + "cuda_time_us": 4.566, + "pct_cuda_time": 0.07912853933846047, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.566, + "pct_cuda_time": 0.07912853933846047, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 372.196, + "cuda_time_us": 46.896, + "pct_cuda_time": 0.8127052082383799, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 58.247, + "cuda_time_us": 17.175, + "pct_cuda_time": 0.29764184475209343, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 17.175, + "pct_cuda_time": 0.29764184475209343, + "trace": "_rocm_C::wvSplitK(bfloat16[6144, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 23.224, + "cuda_time_us": 3.685, + "pct_cuda_time": 0.06386085577359325, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 3.685, + "pct_cuda_time": 0.06386085577359325, + "trace": "_C::rotary_embedding(int64[1], bfloat16[1, 4096], bfloat16[1, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 182.572, + "cuda_time_us": 11.454999999999998, + "pct_cuda_time": 0.19851454623785905, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 2.964, + "pct_cuda_time": 0.05136596377555778, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[1], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_QKV_mfma4_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0, __hip_bfloat16, 16, 128, 256, false, 4>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16 const*, int, float, int const*, int const*, int const*, int, float const*, int, int, int, float*, float*, __hip_bfloat16*, __hip_bfloat16*, int, float const*, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 5.207, + "pct_cuda_time": 0.09023703555308007, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_reduce_kernel<__hip_bfloat16, __hip_bfloat16, 128, 128, 256, 1>(__hip_bfloat16*, float const*, float const*, __hip_bfloat16 const*, int const*, int const*, int, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 3.284, + "pct_cuda_time": 0.05691154690922123, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 66.99, + "cuda_time_us": 14.581, + "pct_cuda_time": 0.252687961474834, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 14.581, + "pct_cuda_time": 0.252687961474834, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 30.355, + "cuda_time_us": 4.927, + "pct_cuda_time": 0.08538465031112455, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.927, + "pct_cuda_time": 0.08538465031112455, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 146.289, + "cuda_time_us": 109.5, + "pct_cuda_time": 1.8976292285504646, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 51.187, + "cuda_time_us": 65.056, + "pct_cuda_time": 1.1274170510737809, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 65.056, + "pct_cuda_time": 1.1274170510737809, + "trace": "_rocm_C::wvSplitK(bfloat16[28672, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 20.641, + "cuda_time_us": 8.371, + "pct_cuda_time": 0.14506898878717753, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 8.371, + "pct_cuda_time": 0.14506898878717753, + "trace": "_C::silu_and_mul(bfloat16[1, 14336], bfloat16[1, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 57.095, + "cuda_time_us": 36.073, + "pct_cuda_time": 0.625143188689506, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 36.073, + "pct_cuda_time": 0.625143188689506, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 14336], bfloat16[1, 14336], 304)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 577.362, + "cuda_time_us": 163.913, + "pct_cuda_time": 2.8406036505880574, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 28.483, + "cuda_time_us": 4.646, + "pct_cuda_time": 0.08051493512187632, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.646, + "pct_cuda_time": 0.08051493512187632, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 341.119, + "cuda_time_us": 48.71600000000001, + "pct_cuda_time": 0.8442457123110909, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 53.379, + "cuda_time_us": 17.344, + "pct_cuda_time": 0.30057060584455947, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 17.344, + "pct_cuda_time": 0.30057060584455947, + "trace": "_rocm_C::wvSplitK(bfloat16[6144, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 24.246, + "cuda_time_us": 4.125, + "pct_cuda_time": 0.07148603258238051, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.125, + "pct_cuda_time": 0.07148603258238051, + "trace": "_C::rotary_embedding(int64[1], bfloat16[1, 4096], bfloat16[1, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 163.815, + "cuda_time_us": 13.697000000000001, + "pct_cuda_time": 0.23736828806808874, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 2.963, + "pct_cuda_time": 0.051348633828265086, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[1], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_QKV_mfma4_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0, __hip_bfloat16, 16, 128, 256, false, 4>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16 const*, int, float, int const*, int const*, int const*, int, float const*, int, int, int, float*, float*, __hip_bfloat16*, __hip_bfloat16*, int, float const*, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 6.889, + "pct_cuda_time": 0.11938600689939864, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_reduce_kernel<__hip_bfloat16, __hip_bfloat16, 128, 128, 256, 1>(__hip_bfloat16*, float const*, float const*, __hip_bfloat16 const*, int const*, int const*, int, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 3.845, + "pct_cuda_time": 0.066633647340425, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 65.348, + "cuda_time_us": 13.55, + "pct_cuda_time": 0.23482078581606203, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 13.55, + "pct_cuda_time": 0.23482078581606203, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 28.212, + "cuda_time_us": 4.766, + "pct_cuda_time": 0.08259452879700013, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.766, + "pct_cuda_time": 0.08259452879700013, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 151.797, + "cuda_time_us": 105.785, + "pct_cuda_time": 1.8332484743580901, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 57.737, + "cuda_time_us": 62.352, + "pct_cuda_time": 1.0805568735943247, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 62.352, + "pct_cuda_time": 1.0805568735943247, + "trace": "_rocm_C::wvSplitK(bfloat16[28672, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 20.841, + "cuda_time_us": 8.211, + "pct_cuda_time": 0.1422961972203458, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 8.211, + "pct_cuda_time": 0.1422961972203458, + "trace": "_C::silu_and_mul(bfloat16[1, 14336], bfloat16[1, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 55.262, + "cuda_time_us": 35.222, + "pct_cuda_time": 0.6103954035434197, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 35.222, + "pct_cuda_time": 0.6103954035434197, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 14336], bfloat16[1, 14336], 304)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 624.823, + "cuda_time_us": 162.452, + "pct_cuda_time": 2.815284597593425, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 28.923, + "cuda_time_us": 4.686, + "pct_cuda_time": 0.08120813301358426, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.686, + "pct_cuda_time": 0.08120813301358426, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 404.243, + "cuda_time_us": 45.943, + "pct_cuda_time": 0.7961897684684383, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 104.927, + "cuda_time_us": 16.904, + "pct_cuda_time": 0.29294542903577214, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 16.904, + "pct_cuda_time": 0.29294542903577214, + "trace": "_rocm_C::wvSplitK(bfloat16[6144, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 25.268, + "cuda_time_us": 3.805, + "pct_cuda_time": 0.06594044944871705, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 3.805, + "pct_cuda_time": 0.06594044944871705, + "trace": "_C::rotary_embedding(int64[1], bfloat16[1, 4096], bfloat16[1, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 175.813, + "cuda_time_us": 10.492999999999999, + "pct_cuda_time": 0.1818431369422833, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 3.004, + "pct_cuda_time": 0.05205916166726571, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[1], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_QKV_mfma4_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0, __hip_bfloat16, 16, 128, 256, false, 4>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16 const*, int, float, int const*, int const*, int const*, int, float const*, int, int, int, float*, float*, __hip_bfloat16*, __hip_bfloat16*, int, float const*, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 4.846, + "pct_cuda_time": 0.08398092458041599, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_reduce_kernel<__hip_bfloat16, __hip_bfloat16, 128, 128, 256, 1>(__hip_bfloat16*, float const*, float const*, __hip_bfloat16 const*, int const*, int const*, int, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 2.643, + "pct_cuda_time": 0.04580305069460162, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 60.771, + "cuda_time_us": 14.741, + "pct_cuda_time": 0.2554607530416657, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 14.741, + "pct_cuda_time": 0.2554607530416657, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 27.892, + "cuda_time_us": 4.646, + "pct_cuda_time": 0.08051493512187632, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.646, + "pct_cuda_time": 0.08051493512187632, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 144.446, + "cuda_time_us": 107.17699999999999, + "pct_cuda_time": 1.8573717609895262, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 51.847, + "cuda_time_us": 63.945, + "pct_cuda_time": 1.1081634796315931, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 63.945, + "pct_cuda_time": 1.1081634796315931, + "trace": "_rocm_C::wvSplitK(bfloat16[28672, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 17.346, + "cuda_time_us": 8.171, + "pct_cuda_time": 0.14160299932863785, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 8.171, + "pct_cuda_time": 0.14160299932863785, + "trace": "_C::silu_and_mul(bfloat16[1, 14336], bfloat16[1, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 54.692, + "cuda_time_us": 35.061, + "pct_cuda_time": 0.6076052820292953, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 35.061, + "pct_cuda_time": 0.6076052820292953, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 14336], bfloat16[1, 14336], 304)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 594.988, + "cuda_time_us": 163.50400000000002, + "pct_cuda_time": 2.8335157021453443, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 26.199, + "cuda_time_us": 4.526, + "pct_cuda_time": 0.07843534144675253, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.526, + "pct_cuda_time": 0.07843534144675253, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 351.425, + "cuda_time_us": 48.406000000000006, + "pct_cuda_time": 0.8388734286503543, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 56.164, + "cuda_time_us": 17.014, + "pct_cuda_time": 0.29485172323796893, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 17.014, + "pct_cuda_time": 0.29485172323796893, + "trace": "_rocm_C::wvSplitK(bfloat16[6144, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 22.253, + "cuda_time_us": 3.925, + "pct_cuda_time": 0.06802004312384086, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 3.925, + "pct_cuda_time": 0.06802004312384086, + "trace": "_C::rotary_embedding(int64[1], bfloat16[1, 4096], bfloat16[1, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 180.499, + "cuda_time_us": 13.017, + "pct_cuda_time": 0.22558392390905385, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 2.683, + "pct_cuda_time": 0.04649624858630955, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[1], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_QKV_mfma4_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0, __hip_bfloat16, 16, 128, 256, false, 4>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16 const*, int, float, int const*, int const*, int const*, int, float const*, int, int, int, float*, float*, __hip_bfloat16*, __hip_bfloat16*, int, float const*, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 6.409, + "pct_cuda_time": 0.11106763219890345, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_reduce_kernel<__hip_bfloat16, __hip_bfloat16, 128, 128, 256, 1>(__hip_bfloat16*, float const*, float const*, __hip_bfloat16 const*, int const*, int const*, int, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 3.925, + "pct_cuda_time": 0.06802004312384086, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 58.217, + "cuda_time_us": 14.45, + "pct_cuda_time": 0.2504177383794905, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 14.45, + "pct_cuda_time": 0.2504177383794905, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 25.318, + "cuda_time_us": 4.606, + "pct_cuda_time": 0.0798217372301684, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.606, + "pct_cuda_time": 0.0798217372301684, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 167.189, + "cuda_time_us": 105.96600000000001, + "pct_cuda_time": 1.8363851948180687, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 60.3, + "cuda_time_us": 61.791, + "pct_cuda_time": 1.0708347731631211, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 61.791, + "pct_cuda_time": 1.0708347731631211, + "trace": "_rocm_C::wvSplitK(bfloat16[28672, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 28.853, + "cuda_time_us": 8.132, + "pct_cuda_time": 0.1409271313842226, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 8.132, + "pct_cuda_time": 0.1409271313842226, + "trace": "_C::silu_and_mul(bfloat16[1, 14336], bfloat16[1, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 61.011, + "cuda_time_us": 36.043, + "pct_cuda_time": 0.624623290270725, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 36.043, + "pct_cuda_time": 0.624623290270725, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 14336], bfloat16[1, 14336], 304)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 578.274, + "cuda_time_us": 164.474, + "pct_cuda_time": 2.8503257510192612, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 26.329, + "cuda_time_us": 4.606, + "pct_cuda_time": 0.0798217372301684, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.606, + "pct_cuda_time": 0.0798217372301684, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 358.485, + "cuda_time_us": 47.94500000000001, + "pct_cuda_time": 0.8308843229484204, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 54.11, + "cuda_time_us": 16.533, + "pct_cuda_time": 0.2865160185901811, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 16.533, + "pct_cuda_time": 0.2865160185901811, + "trace": "_rocm_C::wvSplitK(bfloat16[6144, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 24.537, + "cuda_time_us": 3.725, + "pct_cuda_time": 0.0645540536653012, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 3.725, + "pct_cuda_time": 0.0645540536653012, + "trace": "_C::rotary_embedding(int64[1], bfloat16[1, 4096], bfloat16[1, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 184.165, + "cuda_time_us": 13.417000000000002, + "pct_cuda_time": 0.2325159028261332, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 2.723, + "pct_cuda_time": 0.04718944647801748, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[1], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_QKV_mfma4_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0, __hip_bfloat16, 16, 128, 256, false, 4>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16 const*, int, float, int const*, int const*, int const*, int, float const*, int, int, int, float*, float*, __hip_bfloat16*, __hip_bfloat16*, int, float const*, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 6.769, + "pct_cuda_time": 0.11730641322427483, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_reduce_kernel<__hip_bfloat16, __hip_bfloat16, 128, 128, 256, 1>(__hip_bfloat16*, float const*, float const*, __hip_bfloat16 const*, int const*, int const*, int, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 3.925, + "pct_cuda_time": 0.06802004312384086, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 62.033, + "cuda_time_us": 14.27, + "pct_cuda_time": 0.24729834786680482, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 14.27, + "pct_cuda_time": 0.24729834786680482, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 29.925, + "cuda_time_us": 4.686, + "pct_cuda_time": 0.08120813301358426, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.686, + "pct_cuda_time": 0.08120813301358426, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 143.725, + "cuda_time_us": 107.237, + "pct_cuda_time": 1.8584115578270881, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 52.389, + "cuda_time_us": 63.584, + "pct_cuda_time": 1.1019073686589291, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 63.584, + "pct_cuda_time": 1.1019073686589291, + "trace": "_rocm_C::wvSplitK(bfloat16[28672, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 21.372, + "cuda_time_us": 8.051, + "pct_cuda_time": 0.13952340565351407, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 8.051, + "pct_cuda_time": 0.13952340565351407, + "trace": "_C::silu_and_mul(bfloat16[1, 14336], bfloat16[1, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 53.871, + "cuda_time_us": 35.602, + "pct_cuda_time": 0.616980783514645, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 35.602, + "pct_cuda_time": 0.616980783514645, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 14336], bfloat16[1, 14336], 304)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 578.383, + "cuda_time_us": 163.785, + "pct_cuda_time": 2.838385417334592, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 26.72, + "cuda_time_us": 4.606, + "pct_cuda_time": 0.0798217372301684, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.606, + "pct_cuda_time": 0.0798217372301684, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 355.53, + "cuda_time_us": 47.136, + "pct_cuda_time": 0.8168643955886274, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 51.036, + "cuda_time_us": 17.746, + "pct_cuda_time": 0.3075372446562241, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 17.746, + "pct_cuda_time": 0.3075372446562241, + "trace": "_rocm_C::wvSplitK(bfloat16[6144, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 33.7, + "cuda_time_us": 3.885, + "pct_cuda_time": 0.06732684523213292, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 3.885, + "pct_cuda_time": 0.06732684523213292, + "trace": "_C::rotary_embedding(int64[1], bfloat16[1, 4096], bfloat16[1, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 174.791, + "cuda_time_us": 10.734, + "pct_cuda_time": 0.18601965423982364, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 2.924, + "pct_cuda_time": 0.05067276588384984, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[1], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_QKV_mfma4_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0, __hip_bfloat16, 16, 128, 256, false, 4>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16 const*, int, float, int const*, int const*, int const*, int, float const*, int, int, int, float*, float*, __hip_bfloat16*, __hip_bfloat16*, int, float const*, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 5.006, + "pct_cuda_time": 0.08675371614724772, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_reduce_kernel<__hip_bfloat16, __hip_bfloat16, 128, 128, 256, 1>(__hip_bfloat16*, float const*, float const*, __hip_bfloat16 const*, int const*, int const*, int, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 2.804, + "pct_cuda_time": 0.04859317220872605, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 57.796, + "cuda_time_us": 14.771, + "pct_cuda_time": 0.2559806514604467, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 14.771, + "pct_cuda_time": 0.2559806514604467, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 28.603, + "cuda_time_us": 4.766, + "pct_cuda_time": 0.08259452879700013, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.766, + "pct_cuda_time": 0.08259452879700013, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 146.889, + "cuda_time_us": 107.277, + "pct_cuda_time": 1.8591047557187963, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 52.368, + "cuda_time_us": 64.396, + "pct_cuda_time": 1.1159792858606001, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 64.396, + "pct_cuda_time": 1.1159792858606001, + "trace": "_rocm_C::wvSplitK(bfloat16[28672, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 21.763, + "cuda_time_us": 8.211, + "pct_cuda_time": 0.1422961972203458, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 8.211, + "pct_cuda_time": 0.1422961972203458, + "trace": "_C::silu_and_mul(bfloat16[1, 14336], bfloat16[1, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 56.815, + "cuda_time_us": 34.67, + "pct_cuda_time": 0.6008292726378502, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 34.67, + "pct_cuda_time": 0.6008292726378502, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 14336], bfloat16[1, 14336], 304)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 1225.94, + "cuda_time_us": 162.52200000000002, + "pct_cuda_time": 2.816497693903915, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 28.713, + "cuda_time_us": 4.646, + "pct_cuda_time": 0.08051493512187632, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.646, + "pct_cuda_time": 0.08051493512187632, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 362.961, + "cuda_time_us": 45.903000000000006, + "pct_cuda_time": 0.7954965705767305, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 50.706, + "cuda_time_us": 17.506, + "pct_cuda_time": 0.30337805730597656, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 17.506, + "pct_cuda_time": 0.30337805730597656, + "trace": "_rocm_C::wvSplitK(bfloat16[6144, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 25.137, + "cuda_time_us": 3.764, + "pct_cuda_time": 0.06522992160971643, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 3.764, + "pct_cuda_time": 0.06522992160971643, + "trace": "_C::rotary_embedding(int64[1], bfloat16[1, 4096], bfloat16[1, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 192.718, + "cuda_time_us": 10.933, + "pct_cuda_time": 0.1894683137510706, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 2.963, + "pct_cuda_time": 0.051348633828265086, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[1], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_QKV_mfma4_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0, __hip_bfloat16, 16, 128, 256, false, 4>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16 const*, int, float, int const*, int const*, int const*, int, float const*, int, int, int, float*, float*, __hip_bfloat16*, __hip_bfloat16*, int, float const*, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 4.846, + "pct_cuda_time": 0.08398092458041599, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_reduce_kernel<__hip_bfloat16, __hip_bfloat16, 128, 128, 256, 1>(__hip_bfloat16*, float const*, float const*, __hip_bfloat16 const*, int const*, int const*, int, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 3.124, + "pct_cuda_time": 0.05413875534238951, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 59.138, + "cuda_time_us": 13.7, + "pct_cuda_time": 0.23742027790996678, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 13.7, + "pct_cuda_time": 0.23742027790996678, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 31.296, + "cuda_time_us": 4.686, + "pct_cuda_time": 0.08120813301358426, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.686, + "pct_cuda_time": 0.08120813301358426, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 779.534, + "cuda_time_us": 107.287, + "pct_cuda_time": 1.8592780551917234, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 611.402, + "cuda_time_us": 62.332, + "pct_cuda_time": 1.0802102746484707, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 62.332, + "pct_cuda_time": 1.0802102746484707, + "trace": "_rocm_C::wvSplitK(bfloat16[28672, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 35.062, + "cuda_time_us": 8.291, + "pct_cuda_time": 0.14368259300376165, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 8.291, + "pct_cuda_time": 0.14368259300376165, + "trace": "_C::silu_and_mul(bfloat16[1, 14336], bfloat16[1, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 103.315, + "cuda_time_us": 36.664, + "pct_cuda_time": 0.6353851875394907, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 36.664, + "pct_cuda_time": 0.6353851875394907, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 14336], bfloat16[1, 14336], 304)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 672.984, + "cuda_time_us": 161.672, + "pct_cuda_time": 2.8017672387051205, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 42.423, + "cuda_time_us": 4.686, + "pct_cuda_time": 0.08120813301358426, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.686, + "pct_cuda_time": 0.08120813301358426, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 429.001, + "cuda_time_us": 46.084, + "pct_cuda_time": 0.7986332910367088, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 53.91, + "cuda_time_us": 17.555, + "pct_cuda_time": 0.30422722472331876, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 17.555, + "pct_cuda_time": 0.30422722472331876, + "trace": "_rocm_C::wvSplitK(bfloat16[6144, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 32.098, + "cuda_time_us": 3.885, + "pct_cuda_time": 0.06732684523213292, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 3.885, + "pct_cuda_time": 0.06732684523213292, + "trace": "_C::rotary_embedding(int64[1], bfloat16[1, 4096], bfloat16[1, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 201.611, + "cuda_time_us": 10.493, + "pct_cuda_time": 0.18184313694228335, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 2.683, + "pct_cuda_time": 0.04649624858630955, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[1], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_QKV_mfma4_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0, __hip_bfloat16, 16, 128, 256, false, 4>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16 const*, int, float, int const*, int const*, int const*, int, float const*, int, int, int, float*, float*, __hip_bfloat16*, __hip_bfloat16*, int, float const*, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 5.287, + "pct_cuda_time": 0.09162343133649595, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_reduce_kernel<__hip_bfloat16, __hip_bfloat16, 128, 128, 256, 1>(__hip_bfloat16*, float const*, float const*, __hip_bfloat16 const*, int const*, int const*, int, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 2.523, + "pct_cuda_time": 0.04372345701947783, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 65.818, + "cuda_time_us": 14.151, + "pct_cuda_time": 0.24523608413897371, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 14.151, + "pct_cuda_time": 0.24523608413897371, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 30.175, + "cuda_time_us": 4.686, + "pct_cuda_time": 0.08120813301358426, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.686, + "pct_cuda_time": 0.08120813301358426, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 151.857, + "cuda_time_us": 106.216, + "pct_cuda_time": 1.8407176816412432, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 58.377, + "cuda_time_us": 63.845, + "pct_cuda_time": 1.1064304849023234, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 63.845, + "pct_cuda_time": 1.1064304849023234, + "trace": "_rocm_C::wvSplitK(bfloat16[28672, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 21.802, + "cuda_time_us": 7.931, + "pct_cuda_time": 0.13744381197839026, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 7.931, + "pct_cuda_time": 0.13744381197839026, + "trace": "_C::silu_and_mul(bfloat16[1, 14336], bfloat16[1, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 54.972, + "cuda_time_us": 34.44, + "pct_cuda_time": 0.5968433847605297, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 34.44, + "pct_cuda_time": 0.5968433847605297, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 14336], bfloat16[1, 14336], 304)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 614.047, + "cuda_time_us": 161.632, + "pct_cuda_time": 2.8010740408134125, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 28.052, + "cuda_time_us": 4.526, + "pct_cuda_time": 0.07843534144675253, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.526, + "pct_cuda_time": 0.07843534144675253, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 370.984, + "cuda_time_us": 46.184, + "pct_cuda_time": 0.8003662857659786, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 56.615, + "cuda_time_us": 17.315, + "pct_cuda_time": 0.3000680373730712, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 17.315, + "pct_cuda_time": 0.3000680373730712, + "trace": "_rocm_C::wvSplitK(bfloat16[6144, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 22.924, + "cuda_time_us": 3.525, + "pct_cuda_time": 0.06108806420676153, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 3.525, + "pct_cuda_time": 0.06108806420676153, + "trace": "_C::rotary_embedding(int64[1], bfloat16[1, 4096], bfloat16[1, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 184.856, + "cuda_time_us": 10.373000000000001, + "pct_cuda_time": 0.17976354326715954, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 2.603, + "pct_cuda_time": 0.045109852802893695, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[1], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_QKV_mfma4_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0, __hip_bfloat16, 16, 128, 256, false, 4>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16 const*, int, float, int const*, int const*, int const*, int, float const*, int, int, int, float*, float*, __hip_bfloat16*, __hip_bfloat16*, int, float const*, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 5.207, + "pct_cuda_time": 0.09023703555308007, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_reduce_kernel<__hip_bfloat16, __hip_bfloat16, 128, 128, 256, 1>(__hip_bfloat16*, float const*, float const*, __hip_bfloat16 const*, int const*, int const*, int, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 2.563, + "pct_cuda_time": 0.04441665491118576, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 69.414, + "cuda_time_us": 14.971, + "pct_cuda_time": 0.25944664091898634, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 14.971, + "pct_cuda_time": 0.25944664091898634, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 27.511, + "cuda_time_us": 4.686, + "pct_cuda_time": 0.08120813301358426, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.686, + "pct_cuda_time": 0.08120813301358426, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 165.968, + "cuda_time_us": 106.236, + "pct_cuda_time": 1.8410642805870971, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 62.404, + "cuda_time_us": 63.484, + "pct_cuda_time": 1.1001743739296592, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 63.484, + "pct_cuda_time": 1.1001743739296592, + "trace": "_rocm_C::wvSplitK(bfloat16[28672, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 23.745, + "cuda_time_us": 8.091, + "pct_cuda_time": 0.14021660354522197, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 8.091, + "pct_cuda_time": 0.14021660354522197, + "trace": "_C::silu_and_mul(bfloat16[1, 14336], bfloat16[1, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 62.774, + "cuda_time_us": 34.661, + "pct_cuda_time": 0.600673303112216, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 34.661, + "pct_cuda_time": 0.600673303112216, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 14336], bfloat16[1, 14336], 304)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 587.237, + "cuda_time_us": 168.02, + "pct_cuda_time": 2.9117777441191697, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 36.835, + "cuda_time_us": 4.646, + "pct_cuda_time": 0.08051493512187632, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.646, + "pct_cuda_time": 0.08051493512187632, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 340.739, + "cuda_time_us": 54.055, + "pct_cuda_time": 0.9367703009068069, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 54.412, + "cuda_time_us": 17.745, + "pct_cuda_time": 0.3075199147089315, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 17.745, + "pct_cuda_time": 0.3075199147089315, + "trace": "_rocm_C::wvSplitK(bfloat16[6144, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 24.296, + "cuda_time_us": 4.005, + "pct_cuda_time": 0.06940643890725672, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.005, + "pct_cuda_time": 0.06940643890725672, + "trace": "_C::rotary_embedding(int64[1], bfloat16[1, 4096], bfloat16[1, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 166.939, + "cuda_time_us": 13.458, + "pct_cuda_time": 0.23322643066513382, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 2.884, + "pct_cuda_time": 0.04997956799214191, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[1], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_QKV_mfma4_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0, __hip_bfloat16, 16, 128, 256, false, 4>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16 const*, int, float, int const*, int const*, int const*, int, float const*, int, int, int, float*, float*, __hip_bfloat16*, __hip_bfloat16*, int, float const*, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 6.689, + "pct_cuda_time": 0.11592001744085897, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_reduce_kernel<__hip_bfloat16, __hip_bfloat16, 128, 128, 256, 1>(__hip_bfloat16*, float const*, float const*, __hip_bfloat16 const*, int const*, int const*, int, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 3.885, + "pct_cuda_time": 0.06732684523213292, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 59.889, + "cuda_time_us": 18.847, + "pct_cuda_time": 0.326617516625485, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 18.847, + "pct_cuda_time": 0.326617516625485, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 38.197, + "cuda_time_us": 4.726, + "pct_cuda_time": 0.0819013309052922, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.726, + "pct_cuda_time": 0.0819013309052922, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 153.088, + "cuda_time_us": 104.593, + "pct_cuda_time": 1.8125911771851941, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 59.218, + "cuda_time_us": 62.382, + "pct_cuda_time": 1.0810767720131058, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 62.382, + "pct_cuda_time": 1.0810767720131058, + "trace": "_rocm_C::wvSplitK(bfloat16[28672, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 19.729, + "cuda_time_us": 8.131, + "pct_cuda_time": 0.14090980143692994, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 8.131, + "pct_cuda_time": 0.14090980143692994, + "trace": "_C::silu_and_mul(bfloat16[1, 14336], bfloat16[1, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 58.177, + "cuda_time_us": 34.08, + "pct_cuda_time": 0.5906046037351582, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 34.08, + "pct_cuda_time": 0.5906046037351582, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 14336], bfloat16[1, 14336], 304)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 652.324, + "cuda_time_us": 162.321, + "pct_cuda_time": 2.813014374498082, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 27.942, + "cuda_time_us": 4.686, + "pct_cuda_time": 0.08120813301358426, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.686, + "pct_cuda_time": 0.08120813301358426, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 428.3, + "cuda_time_us": 45.872, + "pct_cuda_time": 0.7949593422106567, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 48.072, + "cuda_time_us": 16.954, + "pct_cuda_time": 0.2938119264004071, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 16.954, + "pct_cuda_time": 0.2938119264004071, + "trace": "_rocm_C::wvSplitK(bfloat16[6144, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 23.224, + "cuda_time_us": 3.764, + "pct_cuda_time": 0.06522992160971643, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 3.764, + "pct_cuda_time": 0.06522992160971643, + "trace": "_C::rotary_embedding(int64[1], bfloat16[1, 4096], bfloat16[1, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 255.912, + "cuda_time_us": 10.814, + "pct_cuda_time": 0.18740605002323948, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 2.763, + "pct_cuda_time": 0.04788264436972542, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[1], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_QKV_mfma4_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0, __hip_bfloat16, 16, 128, 256, false, 4>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16 const*, int, float, int const*, int const*, int const*, int, float const*, int, int, int, float*, float*, __hip_bfloat16*, __hip_bfloat16*, int, float const*, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 5.047, + "pct_cuda_time": 0.08746424398624834, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_reduce_kernel<__hip_bfloat16, __hip_bfloat16, 128, 128, 256, 1>(__hip_bfloat16*, float const*, float const*, __hip_bfloat16 const*, int const*, int const*, int, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 3.004, + "pct_cuda_time": 0.05205916166726571, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 62.073, + "cuda_time_us": 14.34, + "pct_cuda_time": 0.24851144417729368, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 14.34, + "pct_cuda_time": 0.24851144417729368, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 30.826, + "cuda_time_us": 4.726, + "pct_cuda_time": 0.0819013309052922, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.726, + "pct_cuda_time": 0.0819013309052922, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 147.751, + "cuda_time_us": 107.03699999999999, + "pct_cuda_time": 1.8549455683685485, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 51.847, + "cuda_time_us": 62.963, + "pct_cuda_time": 1.0911454713901636, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 62.963, + "pct_cuda_time": 1.0911454713901636, + "trace": "_rocm_C::wvSplitK(bfloat16[28672, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 25.108, + "cuda_time_us": 8.131, + "pct_cuda_time": 0.14090980143692994, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 8.131, + "pct_cuda_time": 0.14090980143692994, + "trace": "_C::silu_and_mul(bfloat16[1, 14336], bfloat16[1, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 54.452, + "cuda_time_us": 35.943, + "pct_cuda_time": 0.6228902955414551, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 35.943, + "pct_cuda_time": 0.6228902955414551, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 14336], bfloat16[1, 14336], 304)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 548.849, + "cuda_time_us": 164.243, + "pct_cuda_time": 2.846322533194648, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 27.521, + "cuda_time_us": 4.686, + "pct_cuda_time": 0.08120813301358426, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.686, + "pct_cuda_time": 0.08120813301358426, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 336.742, + "cuda_time_us": 49.166999999999994, + "pct_cuda_time": 0.8520615185400975, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 50.015, + "cuda_time_us": 17.605, + "pct_cuda_time": 0.3050937220879537, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 17.605, + "pct_cuda_time": 0.3050937220879537, + "trace": "_rocm_C::wvSplitK(bfloat16[6144, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 24.476, + "cuda_time_us": 4.045, + "pct_cuda_time": 0.07009963679896465, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.045, + "pct_cuda_time": 0.07009963679896465, + "trace": "_C::rotary_embedding(int64[1], bfloat16[1, 4096], bfloat16[1, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 167.16, + "cuda_time_us": 13.577, + "pct_cuda_time": 0.2352886943929649, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 2.883, + "pct_cuda_time": 0.04996223804484922, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[1], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_QKV_mfma4_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0, __hip_bfloat16, 16, 128, 256, false, 4>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16 const*, int, float, int const*, int const*, int const*, int, float const*, int, int, int, float*, float*, __hip_bfloat16*, __hip_bfloat16*, int, float const*, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 6.849, + "pct_cuda_time": 0.11869280900769069, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_reduce_kernel<__hip_bfloat16, __hip_bfloat16, 128, 128, 256, 1>(__hip_bfloat16*, float const*, float const*, __hip_bfloat16 const*, int const*, int const*, int, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 3.845, + "pct_cuda_time": 0.066633647340425, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 58.678, + "cuda_time_us": 13.94, + "pct_cuda_time": 0.2415794652602144, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 13.94, + "pct_cuda_time": 0.2415794652602144, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 27.111, + "cuda_time_us": 4.686, + "pct_cuda_time": 0.08120813301358426, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.686, + "pct_cuda_time": 0.08120813301358426, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 138.887, + "cuda_time_us": 105.70400000000001, + "pct_cuda_time": 1.831844748627382, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 50.886, + "cuda_time_us": 62.382, + "pct_cuda_time": 1.0810767720131058, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 62.382, + "pct_cuda_time": 1.0810767720131058, + "trace": "_rocm_C::wvSplitK(bfloat16[28672, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 22.073, + "cuda_time_us": 8.091, + "pct_cuda_time": 0.14021660354522197, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 8.091, + "pct_cuda_time": 0.14021660354522197, + "trace": "_C::silu_and_mul(bfloat16[1, 14336], bfloat16[1, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 52.038, + "cuda_time_us": 35.231, + "pct_cuda_time": 0.610551373069054, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 35.231, + "pct_cuda_time": 0.610551373069054, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 14336], bfloat16[1, 14336], 304)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 576.481, + "cuda_time_us": 167.39, + "pct_cuda_time": 2.9008598773247694, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 26.87, + "cuda_time_us": 4.686, + "pct_cuda_time": 0.08120813301358426, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.686, + "pct_cuda_time": 0.08120813301358426, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 356.963, + "cuda_time_us": 51.732, + "pct_cuda_time": 0.8965128333458687, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 51.557, + "cuda_time_us": 18.166, + "pct_cuda_time": 0.31481582251915746, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 18.166, + "pct_cuda_time": 0.31481582251915746, + "trace": "_rocm_C::wvSplitK(bfloat16[6144, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 27.421, + "cuda_time_us": 3.725, + "pct_cuda_time": 0.0645540536653012, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 3.725, + "pct_cuda_time": 0.0645540536653012, + "trace": "_C::rotary_embedding(int64[1], bfloat16[1, 4096], bfloat16[1, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 182.803, + "cuda_time_us": 15.861, + "pct_cuda_time": 0.2748702940094878, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 5.247, + "pct_cuda_time": 0.09093023344478801, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[1], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_QKV_mfma4_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0, __hip_bfloat16, 16, 128, 256, false, 4>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16 const*, int, float, int const*, int const*, int const*, int, float const*, int, int, int, float*, float*, __hip_bfloat16*, __hip_bfloat16*, int, float const*, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 6.689, + "pct_cuda_time": 0.11592001744085897, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_reduce_kernel<__hip_bfloat16, __hip_bfloat16, 128, 128, 256, 1>(__hip_bfloat16*, float const*, float const*, __hip_bfloat16 const*, int const*, int const*, int, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 3.925, + "pct_cuda_time": 0.06802004312384086, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 59.999, + "cuda_time_us": 13.98, + "pct_cuda_time": 0.24227266315192234, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 13.98, + "pct_cuda_time": 0.24227266315192234, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 29.414, + "cuda_time_us": 4.606, + "pct_cuda_time": 0.0798217372301684, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.606, + "pct_cuda_time": 0.0798217372301684, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 144.937, + "cuda_time_us": 106.366, + "pct_cuda_time": 1.843317173735148, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 50.716, + "cuda_time_us": 63.093, + "pct_cuda_time": 1.0933983645382144, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 63.093, + "pct_cuda_time": 1.0933983645382144, + "trace": "_rocm_C::wvSplitK(bfloat16[28672, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 26.249, + "cuda_time_us": 8.011, + "pct_cuda_time": 0.13883020776180613, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 8.011, + "pct_cuda_time": 0.13883020776180613, + "trace": "_C::silu_and_mul(bfloat16[1, 14336], bfloat16[1, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 51.347, + "cuda_time_us": 35.262, + "pct_cuda_time": 0.6110886014351277, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 35.262, + "pct_cuda_time": 0.6110886014351277, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 14336], bfloat16[1, 14336], 304)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 572.775, + "cuda_time_us": 158.074, + "pct_cuda_time": 2.7394140883459923, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 32.528, + "cuda_time_us": 4.526, + "pct_cuda_time": 0.07843534144675253, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.526, + "pct_cuda_time": 0.07843534144675253, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 347.238, + "cuda_time_us": 44.981, + "pct_cuda_time": 0.7795183591728625, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 51.266, + "cuda_time_us": 16.623, + "pct_cuda_time": 0.288075713846524, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 16.623, + "pct_cuda_time": 0.288075713846524, + "trace": "_rocm_C::wvSplitK(bfloat16[6144, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 25.538, + "cuda_time_us": 3.844, + "pct_cuda_time": 0.06661631739313228, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 3.844, + "pct_cuda_time": 0.06661631739313228, + "trace": "_C::rotary_embedding(int64[1], bfloat16[1, 4096], bfloat16[1, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 181.03, + "cuda_time_us": 10.574, + "pct_cuda_time": 0.1832468626729919, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 2.563, + "pct_cuda_time": 0.04441665491118576, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[1], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_QKV_mfma4_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0, __hip_bfloat16, 16, 128, 256, false, 4>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16 const*, int, float, int const*, int const*, int const*, int, float const*, int, int, int, float*, float*, __hip_bfloat16*, __hip_bfloat16*, int, float const*, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 5.568, + "pct_cuda_time": 0.09649314652574417, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_reduce_kernel<__hip_bfloat16, __hip_bfloat16, 128, 128, 256, 1>(__hip_bfloat16*, float const*, float const*, __hip_bfloat16 const*, int const*, int const*, int, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 2.443, + "pct_cuda_time": 0.042337061236061964, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 56.985, + "cuda_time_us": 13.94, + "pct_cuda_time": 0.2415794652602144, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 13.94, + "pct_cuda_time": 0.2415794652602144, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 28.983, + "cuda_time_us": 4.726, + "pct_cuda_time": 0.0819013309052922, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.726, + "pct_cuda_time": 0.0819013309052922, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 146.359, + "cuda_time_us": 103.84100000000001, + "pct_cuda_time": 1.7995590568210849, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 57.125, + "cuda_time_us": 62.221, + "pct_cuda_time": 1.0782866504989812, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 62.221, + "pct_cuda_time": 1.0782866504989812, + "trace": "_rocm_C::wvSplitK(bfloat16[28672, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 20.46, + "cuda_time_us": 8.011, + "pct_cuda_time": 0.13883020776180613, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 8.011, + "pct_cuda_time": 0.13883020776180613, + "trace": "_C::silu_and_mul(bfloat16[1, 14336], bfloat16[1, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 52.207, + "cuda_time_us": 33.609, + "pct_cuda_time": 0.5824421985602974, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 33.609, + "pct_cuda_time": 0.5824421985602974, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 14336], bfloat16[1, 14336], 304)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 561.188, + "cuda_time_us": 163.56400000000002, + "pct_cuda_time": 2.834555498982906, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 25.408, + "cuda_time_us": 4.606, + "pct_cuda_time": 0.0798217372301684, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.606, + "pct_cuda_time": 0.0798217372301684, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 338.755, + "cuda_time_us": 48.177, + "pct_cuda_time": 0.8349048707203263, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 51.657, + "cuda_time_us": 16.945, + "pct_cuda_time": 0.2936559568747728, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 16.945, + "pct_cuda_time": 0.2936559568747728, + "trace": "_rocm_C::wvSplitK(bfloat16[6144, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 24.226, + "cuda_time_us": 3.725, + "pct_cuda_time": 0.0645540536653012, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 3.725, + "pct_cuda_time": 0.0645540536653012, + "trace": "_C::rotary_embedding(int64[1], bfloat16[1, 4096], bfloat16[1, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 170.895, + "cuda_time_us": 13.177, + "pct_cuda_time": 0.22835671547588557, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 2.683, + "pct_cuda_time": 0.04649624858630955, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[1], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_QKV_mfma4_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0, __hip_bfloat16, 16, 128, 256, false, 4>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16 const*, int, float, int const*, int const*, int const*, int, float const*, int, int, int, float*, float*, __hip_bfloat16*, __hip_bfloat16*, int, float const*, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 6.569, + "pct_cuda_time": 0.11384042376573517, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_reduce_kernel<__hip_bfloat16, __hip_bfloat16, 128, 128, 256, 1>(__hip_bfloat16*, float const*, float const*, __hip_bfloat16 const*, int const*, int const*, int, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 3.925, + "pct_cuda_time": 0.06802004312384086, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 61.642, + "cuda_time_us": 14.33, + "pct_cuda_time": 0.24833814470436674, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 14.33, + "pct_cuda_time": 0.24833814470436674, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 29.534, + "cuda_time_us": 4.726, + "pct_cuda_time": 0.0819013309052922, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.726, + "pct_cuda_time": 0.0819013309052922, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 150.084, + "cuda_time_us": 106.055, + "pct_cuda_time": 1.837927560127119, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 56.765, + "cuda_time_us": 64.004, + "pct_cuda_time": 1.1091859465218625, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 64.004, + "pct_cuda_time": 1.1091859465218625, + "trace": "_rocm_C::wvSplitK(bfloat16[28672, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 22.513, + "cuda_time_us": 8.252, + "pct_cuda_time": 0.14300672505934642, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 8.252, + "pct_cuda_time": 0.14300672505934642, + "trace": "_C::silu_and_mul(bfloat16[1, 14336], bfloat16[1, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 51.587, + "cuda_time_us": 33.799, + "pct_cuda_time": 0.58573488854591, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 33.799, + "pct_cuda_time": 0.58573488854591, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 14336], bfloat16[1, 14336], 304)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 553.426, + "cuda_time_us": 163.601, + "pct_cuda_time": 2.8351967070327357, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 26.89, + "cuda_time_us": 4.565, + "pct_cuda_time": 0.07911120939116778, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.565, + "pct_cuda_time": 0.07911120939116778, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 340.408, + "cuda_time_us": 50.408, + "pct_cuda_time": 0.8735679831303362, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 53.079, + "cuda_time_us": 20.229, + "pct_cuda_time": 0.35056750378399404, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 20.229, + "pct_cuda_time": 0.35056750378399404, + "trace": "_rocm_C::wvSplitK(bfloat16[6144, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 24.216, + "cuda_time_us": 3.965, + "pct_cuda_time": 0.06871324101554878, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 3.965, + "pct_cuda_time": 0.06871324101554878, + "trace": "_C::rotary_embedding(int64[1], bfloat16[1, 4096], bfloat16[1, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 171.246, + "cuda_time_us": 12.214, + "pct_cuda_time": 0.21166797623301709, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 2.963, + "pct_cuda_time": 0.051348633828265086, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[1], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_QKV_mfma4_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0, __hip_bfloat16, 16, 128, 256, false, 4>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16 const*, int, float, int const*, int const*, int const*, int, float const*, int, int, int, float*, float*, __hip_bfloat16*, __hip_bfloat16*, int, float const*, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 5.367, + "pct_cuda_time": 0.09300982711991182, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_reduce_kernel<__hip_bfloat16, __hip_bfloat16, 128, 128, 256, 1>(__hip_bfloat16*, float const*, float const*, __hip_bfloat16 const*, int const*, int const*, int, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 3.884, + "pct_cuda_time": 0.06730951528484022, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 57.476, + "cuda_time_us": 14.0, + "pct_cuda_time": 0.2426192620977763, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 14.0, + "pct_cuda_time": 0.2426192620977763, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 27.692, + "cuda_time_us": 4.726, + "pct_cuda_time": 0.0819013309052922, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.726, + "pct_cuda_time": 0.0819013309052922, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 141.18, + "cuda_time_us": 103.90199999999999, + "pct_cuda_time": 1.8006161836059391, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 51.878, + "cuda_time_us": 62.202, + "pct_cuda_time": 1.07795738150042, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 62.202, + "pct_cuda_time": 1.07795738150042, + "trace": "_rocm_C::wvSplitK(bfloat16[28672, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 19.91, + "cuda_time_us": 8.091, + "pct_cuda_time": 0.14021660354522197, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 8.091, + "pct_cuda_time": 0.14021660354522197, + "trace": "_C::silu_and_mul(bfloat16[1, 14336], bfloat16[1, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 52.248, + "cuda_time_us": 33.609, + "pct_cuda_time": 0.5824421985602974, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 33.609, + "pct_cuda_time": 0.5824421985602974, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 14336], bfloat16[1, 14336], 304)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 536.14, + "cuda_time_us": 165.757, + "pct_cuda_time": 2.8725600733957934, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 24.347, + "cuda_time_us": 4.646, + "pct_cuda_time": 0.08051493512187632, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.646, + "pct_cuda_time": 0.08051493512187632, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 312.527, + "cuda_time_us": 48.959, + "pct_cuda_time": 0.8484568895032164, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 49.624, + "cuda_time_us": 17.265, + "pct_cuda_time": 0.29920154000843624, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 17.265, + "pct_cuda_time": 0.29920154000843624, + "trace": "_rocm_C::wvSplitK(bfloat16[6144, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 22.624, + "cuda_time_us": 3.725, + "pct_cuda_time": 0.0645540536653012, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 3.725, + "pct_cuda_time": 0.0645540536653012, + "trace": "_C::rotary_embedding(int64[1], bfloat16[1, 4096], bfloat16[1, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 151.036, + "cuda_time_us": 13.818, + "pct_cuda_time": 0.23946521169050516, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 2.683, + "pct_cuda_time": 0.04649624858630955, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[1], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_QKV_mfma4_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0, __hip_bfloat16, 16, 128, 256, false, 4>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16 const*, int, float, int const*, int const*, int const*, int, float const*, int, int, int, float*, float*, __hip_bfloat16*, __hip_bfloat16*, int, float const*, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 7.17, + "pct_cuda_time": 0.12425572208864684, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_reduce_kernel<__hip_bfloat16, __hip_bfloat16, 128, 128, 256, 1>(__hip_bfloat16*, float const*, float const*, __hip_bfloat16 const*, int const*, int const*, int, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 3.965, + "pct_cuda_time": 0.06871324101554878, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 58.037, + "cuda_time_us": 14.151, + "pct_cuda_time": 0.24523608413897371, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 14.151, + "pct_cuda_time": 0.24523608413897371, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 27.571, + "cuda_time_us": 5.206, + "pct_cuda_time": 0.09021970560578739, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 5.206, + "pct_cuda_time": 0.09021970560578739, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 153.339, + "cuda_time_us": 106.946, + "pct_cuda_time": 1.8533685431649132, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 53.87, + "cuda_time_us": 64.455, + "pct_cuda_time": 1.1170017527508693, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 64.455, + "pct_cuda_time": 1.1170017527508693, + "trace": "_rocm_C::wvSplitK(bfloat16[28672, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 22.584, + "cuda_time_us": 8.171, + "pct_cuda_time": 0.14160299932863785, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 8.171, + "pct_cuda_time": 0.14160299932863785, + "trace": "_C::silu_and_mul(bfloat16[1, 14336], bfloat16[1, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 57.556, + "cuda_time_us": 34.32, + "pct_cuda_time": 0.5947637910854059, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 34.32, + "pct_cuda_time": 0.5947637910854059, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 14336], bfloat16[1, 14336], 304)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 560.937, + "cuda_time_us": 160.32799999999997, + "pct_cuda_time": 2.7784757895437338, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 26.019, + "cuda_time_us": 4.686, + "pct_cuda_time": 0.08120813301358426, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.686, + "pct_cuda_time": 0.08120813301358426, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 339.456, + "cuda_time_us": 46.393, + "pct_cuda_time": 0.8039882447501525, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 49.354, + "cuda_time_us": 16.474, + "pct_cuda_time": 0.2854935516999119, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 16.474, + "pct_cuda_time": 0.2854935516999119, + "trace": "_rocm_C::wvSplitK(bfloat16[6144, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 25.648, + "cuda_time_us": 3.884, + "pct_cuda_time": 0.06730951528484022, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 3.884, + "pct_cuda_time": 0.06730951528484022, + "trace": "_C::rotary_embedding(int64[1], bfloat16[1, 4096], bfloat16[1, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 172.287, + "cuda_time_us": 11.614, + "pct_cuda_time": 0.20127000785739815, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 2.683, + "pct_cuda_time": 0.04649624858630955, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[1], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_QKV_mfma4_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0, __hip_bfloat16, 16, 128, 256, false, 4>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16 const*, int, float, int const*, int const*, int const*, int, float const*, int, int, int, float*, float*, __hip_bfloat16*, __hip_bfloat16*, int, float const*, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 5.367, + "pct_cuda_time": 0.09300982711991182, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_reduce_kernel<__hip_bfloat16, __hip_bfloat16, 128, 128, 256, 1>(__hip_bfloat16*, float const*, float const*, __hip_bfloat16 const*, int const*, int const*, int, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 3.564, + "pct_cuda_time": 0.061763932151176766, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 61.392, + "cuda_time_us": 14.421, + "pct_cuda_time": 0.24991516990800228, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 14.421, + "pct_cuda_time": 0.24991516990800228, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 29.795, + "cuda_time_us": 4.686, + "pct_cuda_time": 0.08120813301358426, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.686, + "pct_cuda_time": 0.08120813301358426, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 147.46, + "cuda_time_us": 104.56299999999999, + "pct_cuda_time": 1.8120712787664128, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 57.146, + "cuda_time_us": 62.593, + "pct_cuda_time": 1.084733390891865, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 62.593, + "pct_cuda_time": 1.084733390891865, + "trace": "_rocm_C::wvSplitK(bfloat16[28672, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 21.102, + "cuda_time_us": 8.091, + "pct_cuda_time": 0.14021660354522197, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 8.091, + "pct_cuda_time": 0.14021660354522197, + "trace": "_C::silu_and_mul(bfloat16[1, 14336], bfloat16[1, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 51.667, + "cuda_time_us": 33.879, + "pct_cuda_time": 0.5871212843293259, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 33.879, + "pct_cuda_time": 0.5871212843293259, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 14336], bfloat16[1, 14336], 304)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 639.845, + "cuda_time_us": 161.59199999999998, + "pct_cuda_time": 2.8003808429217045, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 27.972, + "cuda_time_us": 4.566, + "pct_cuda_time": 0.07912853933846047, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.566, + "pct_cuda_time": 0.07912853933846047, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 355.611, + "cuda_time_us": 48.226, + "pct_cuda_time": 0.8357540381376686, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 49.384, + "cuda_time_us": 17.235, + "pct_cuda_time": 0.2986816415896553, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 17.235, + "pct_cuda_time": 0.2986816415896553, + "trace": "_rocm_C::wvSplitK(bfloat16[6144, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 25.438, + "cuda_time_us": 3.565, + "pct_cuda_time": 0.06178126209846946, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 3.565, + "pct_cuda_time": 0.06178126209846946, + "trace": "_C::rotary_embedding(int64[1], bfloat16[1, 4096], bfloat16[1, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 185.617, + "cuda_time_us": 13.216, + "pct_cuda_time": 0.2290325834203008, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 2.603, + "pct_cuda_time": 0.045109852802893695, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[1], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_QKV_mfma4_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0, __hip_bfloat16, 16, 128, 256, false, 4>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16 const*, int, float, int const*, int const*, int const*, int, float const*, int, int, int, float*, float*, __hip_bfloat16*, __hip_bfloat16*, int, float const*, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 6.648, + "pct_cuda_time": 0.11520948960185834, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_reduce_kernel<__hip_bfloat16, __hip_bfloat16, 128, 128, 256, 1>(__hip_bfloat16*, float const*, float const*, __hip_bfloat16 const*, int const*, int const*, int, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 3.965, + "pct_cuda_time": 0.06871324101554878, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 57.646, + "cuda_time_us": 14.21, + "pct_cuda_time": 0.24625855102924296, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 14.21, + "pct_cuda_time": 0.24625855102924296, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 98.077, + "cuda_time_us": 4.807, + "pct_cuda_time": 0.08330505663600077, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.807, + "pct_cuda_time": 0.08330505663600077, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 140.6, + "cuda_time_us": 103.993, + "pct_cuda_time": 1.8021932088095751, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 59.409, + "cuda_time_us": 62.973, + "pct_cuda_time": 1.0913187708630903, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 62.973, + "pct_cuda_time": 1.0913187708630903, + "trace": "_rocm_C::wvSplitK(bfloat16[28672, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 17.977, + "cuda_time_us": 8.092, + "pct_cuda_time": 0.1402339334925147, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 8.092, + "pct_cuda_time": 0.1402339334925147, + "trace": "_C::silu_and_mul(bfloat16[1, 14336], bfloat16[1, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 47.871, + "cuda_time_us": 32.928, + "pct_cuda_time": 0.5706405044539697, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 32.928, + "pct_cuda_time": 0.5706405044539697, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 14336], bfloat16[1, 14336], 304)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 571.974, + "cuda_time_us": 162.543, + "pct_cuda_time": 2.816861622797061, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 29.974, + "cuda_time_us": 4.686, + "pct_cuda_time": 0.08120813301358426, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.686, + "pct_cuda_time": 0.08120813301358426, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 354.399, + "cuda_time_us": 49.469, + "pct_cuda_time": 0.8572951626224925, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 72.258, + "cuda_time_us": 17.085, + "pct_cuda_time": 0.29608214949575057, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 17.085, + "pct_cuda_time": 0.29608214949575057, + "trace": "_rocm_C::wvSplitK(bfloat16[6144, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 33.63, + "cuda_time_us": 3.965, + "pct_cuda_time": 0.06871324101554878, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 3.965, + "pct_cuda_time": 0.06871324101554878, + "trace": "_C::rotary_embedding(int64[1], bfloat16[1, 4096], bfloat16[1, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 153.099, + "cuda_time_us": 13.618, + "pct_cuda_time": 0.23599922223196554, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 2.603, + "pct_cuda_time": 0.045109852802893695, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[1], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_QKV_mfma4_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0, __hip_bfloat16, 16, 128, 256, false, 4>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16 const*, int, float, int const*, int const*, int const*, int, float const*, int, int, int, float*, float*, __hip_bfloat16*, __hip_bfloat16*, int, float const*, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 7.17, + "pct_cuda_time": 0.12425572208864684, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_reduce_kernel<__hip_bfloat16, __hip_bfloat16, 128, 128, 256, 1>(__hip_bfloat16*, float const*, float const*, __hip_bfloat16 const*, int const*, int const*, int, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 3.845, + "pct_cuda_time": 0.066633647340425, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 60.841, + "cuda_time_us": 14.801, + "pct_cuda_time": 0.2565005498792276, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 14.801, + "pct_cuda_time": 0.2565005498792276, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 26.169, + "cuda_time_us": 4.726, + "pct_cuda_time": 0.0819013309052922, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.726, + "pct_cuda_time": 0.0819013309052922, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 140.98, + "cuda_time_us": 103.662, + "pct_cuda_time": 1.796456996255692, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 55.984, + "cuda_time_us": 61.962, + "pct_cuda_time": 1.0737981941501724, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 61.962, + "pct_cuda_time": 1.0737981941501724, + "trace": "_rocm_C::wvSplitK(bfloat16[28672, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 19.89, + "cuda_time_us": 8.051, + "pct_cuda_time": 0.13952340565351407, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 8.051, + "pct_cuda_time": 0.13952340565351407, + "trace": "_C::silu_and_mul(bfloat16[1, 14336], bfloat16[1, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 49.854, + "cuda_time_us": 33.649, + "pct_cuda_time": 0.5831353964520053, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 33.649, + "pct_cuda_time": 0.5831353964520053, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 14336], bfloat16[1, 14336], 304)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 555.539, + "cuda_time_us": 164.094, + "pct_cuda_time": 2.843740371048036, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 26.379, + "cuda_time_us": 4.726, + "pct_cuda_time": 0.0819013309052922, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.726, + "pct_cuda_time": 0.0819013309052922, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 338.615, + "cuda_time_us": 48.296, + "pct_cuda_time": 0.8369671344481574, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 48.072, + "cuda_time_us": 17.255, + "pct_cuda_time": 0.29902824053550925, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 17.255, + "pct_cuda_time": 0.29902824053550925, + "trace": "_rocm_C::wvSplitK(bfloat16[6144, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 25.068, + "cuda_time_us": 3.804, + "pct_cuda_time": 0.06592311950142435, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 3.804, + "pct_cuda_time": 0.06592311950142435, + "trace": "_C::rotary_embedding(int64[1], bfloat16[1, 4096], bfloat16[1, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 177.565, + "cuda_time_us": 12.535999999999998, + "pct_cuda_time": 0.21724821926126592, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 2.803, + "pct_cuda_time": 0.048575842261433355, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[1], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_QKV_mfma4_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0, __hip_bfloat16, 16, 128, 256, false, 4>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16 const*, int, float, int const*, int const*, int const*, int, float const*, int, int, int, float*, float*, __hip_bfloat16*, __hip_bfloat16*, int, float const*, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 5.728, + "pct_cuda_time": 0.09926593809257589, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_reduce_kernel<__hip_bfloat16, __hip_bfloat16, 128, 128, 256, 1>(__hip_bfloat16*, float const*, float const*, __hip_bfloat16 const*, int const*, int const*, int, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 4.005, + "pct_cuda_time": 0.06940643890725672, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 58.387, + "cuda_time_us": 14.701, + "pct_cuda_time": 0.2547675551499578, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 14.701, + "pct_cuda_time": 0.2547675551499578, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 30.155, + "cuda_time_us": 4.927, + "pct_cuda_time": 0.08538465031112455, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.927, + "pct_cuda_time": 0.08538465031112455, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 143.695, + "cuda_time_us": 106.14500000000001, + "pct_cuda_time": 1.839487255383462, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 55.953, + "cuda_time_us": 62.372, + "pct_cuda_time": 1.0809034725401787, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 62.372, + "pct_cuda_time": 1.0809034725401787, + "trace": "_rocm_C::wvSplitK(bfloat16[28672, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 21.732, + "cuda_time_us": 8.171, + "pct_cuda_time": 0.14160299932863785, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 8.171, + "pct_cuda_time": 0.14160299932863785, + "trace": "_C::silu_and_mul(bfloat16[1, 14336], bfloat16[1, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 51.056, + "cuda_time_us": 35.602, + "pct_cuda_time": 0.616980783514645, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 35.602, + "pct_cuda_time": 0.616980783514645, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 14336], bfloat16[1, 14336], 304)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 544.884, + "cuda_time_us": 164.652, + "pct_cuda_time": 2.8534104816373613, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 27.281, + "cuda_time_us": 4.645, + "pct_cuda_time": 0.08049760517458363, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.645, + "pct_cuda_time": 0.08049760517458363, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 324.104, + "cuda_time_us": 49.777, + "pct_cuda_time": 0.8626327863886436, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 48.422, + "cuda_time_us": 17.735, + "pct_cuda_time": 0.3073466152360045, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 17.735, + "pct_cuda_time": 0.3073466152360045, + "trace": "_rocm_C::wvSplitK(bfloat16[6144, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 29.775, + "cuda_time_us": 4.045, + "pct_cuda_time": 0.07009963679896465, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.045, + "pct_cuda_time": 0.07009963679896465, + "trace": "_C::rotary_embedding(int64[1], bfloat16[1, 4096], bfloat16[1, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 162.833, + "cuda_time_us": 13.737, + "pct_cuda_time": 0.23806148595979662, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 2.723, + "pct_cuda_time": 0.04718944647801748, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[1], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_QKV_mfma4_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0, __hip_bfloat16, 16, 128, 256, false, 4>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16 const*, int, float, int const*, int const*, int const*, int, float const*, int, int, int, float*, float*, __hip_bfloat16*, __hip_bfloat16*, int, float const*, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 7.169, + "pct_cuda_time": 0.12423839214135415, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_reduce_kernel<__hip_bfloat16, __hip_bfloat16, 128, 128, 256, 1>(__hip_bfloat16*, float const*, float const*, __hip_bfloat16 const*, int const*, int const*, int, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 3.845, + "pct_cuda_time": 0.066633647340425, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 55.392, + "cuda_time_us": 14.26, + "pct_cuda_time": 0.24712504839387783, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 14.26, + "pct_cuda_time": 0.24712504839387783, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 28.793, + "cuda_time_us": 4.686, + "pct_cuda_time": 0.08120813301358426, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.686, + "pct_cuda_time": 0.08120813301358426, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 145.287, + "cuda_time_us": 105.544, + "pct_cuda_time": 1.82907195706055, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 58.157, + "cuda_time_us": 62.242, + "pct_cuda_time": 1.078650579392128, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 62.242, + "pct_cuda_time": 1.078650579392128, + "trace": "_rocm_C::wvSplitK(bfloat16[28672, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 21.021, + "cuda_time_us": 8.091, + "pct_cuda_time": 0.14021660354522197, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 8.091, + "pct_cuda_time": 0.14021660354522197, + "trace": "_C::silu_and_mul(bfloat16[1, 14336], bfloat16[1, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 50.515, + "cuda_time_us": 35.211, + "pct_cuda_time": 0.6102047741232001, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 35.211, + "pct_cuda_time": 0.6102047741232001, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 14336], bfloat16[1, 14336], 304)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 549.931, + "cuda_time_us": 164.404, + "pct_cuda_time": 2.8491126547087724, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 25.188, + "cuda_time_us": 4.606, + "pct_cuda_time": 0.0798217372301684, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.606, + "pct_cuda_time": 0.0798217372301684, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 338.525, + "cuda_time_us": 48.537, + "pct_cuda_time": 0.8411436517456976, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 46.89, + "cuda_time_us": 17.084, + "pct_cuda_time": 0.2960648195484578, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 17.084, + "pct_cuda_time": 0.2960648195484578, + "trace": "_rocm_C::wvSplitK(bfloat16[6144, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 26.53, + "cuda_time_us": 3.725, + "pct_cuda_time": 0.0645540536653012, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 3.725, + "pct_cuda_time": 0.0645540536653012, + "trace": "_C::rotary_embedding(int64[1], bfloat16[1, 4096], bfloat16[1, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 166.409, + "cuda_time_us": 13.337, + "pct_cuda_time": 0.23112950704271729, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 2.724, + "pct_cuda_time": 0.04720677642531019, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[1], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_QKV_mfma4_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0, __hip_bfloat16, 16, 128, 256, false, 4>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16 const*, int, float, int const*, int const*, int const*, int, float const*, int, int, int, float*, float*, __hip_bfloat16*, __hip_bfloat16*, int, float const*, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 6.648, + "pct_cuda_time": 0.11520948960185834, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_reduce_kernel<__hip_bfloat16, __hip_bfloat16, 128, 128, 256, 1>(__hip_bfloat16*, float const*, float const*, __hip_bfloat16 const*, int const*, int const*, int, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 3.965, + "pct_cuda_time": 0.06871324101554878, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 65.537, + "cuda_time_us": 14.391, + "pct_cuda_time": 0.24939527148922133, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 14.391, + "pct_cuda_time": 0.24939527148922133, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 28.753, + "cuda_time_us": 4.525, + "pct_cuda_time": 0.07841801149945984, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.525, + "pct_cuda_time": 0.07841801149945984, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 141.161, + "cuda_time_us": 106.736, + "pct_cuda_time": 1.8497292542334465, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 53.731, + "cuda_time_us": 63.393, + "pct_cuda_time": 1.0985973487260237, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 63.393, + "pct_cuda_time": 1.0985973487260237, + "trace": "_rocm_C::wvSplitK(bfloat16[28672, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 19.53, + "cuda_time_us": 8.051, + "pct_cuda_time": 0.13952340565351407, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 8.051, + "pct_cuda_time": 0.13952340565351407, + "trace": "_C::silu_and_mul(bfloat16[1, 14336], bfloat16[1, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 51.307, + "cuda_time_us": 35.292, + "pct_cuda_time": 0.6116084998539086, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 35.292, + "pct_cuda_time": 0.6116084998539086, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 14336], bfloat16[1, 14336], 304)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 545.945, + "cuda_time_us": 158.365, + "pct_cuda_time": 2.7444571030081675, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 24.857, + "cuda_time_us": 4.365, + "pct_cuda_time": 0.07564521993262811, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.365, + "pct_cuda_time": 0.07564521993262811, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 326.908, + "cuda_time_us": 44.160000000000004, + "pct_cuda_time": 0.7652904724455573, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 48.933, + "cuda_time_us": 15.592, + "pct_cuda_time": 0.270208538187752, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 15.592, + "pct_cuda_time": 0.270208538187752, + "trace": "_rocm_C::wvSplitK(bfloat16[6144, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 24.797, + "cuda_time_us": 3.805, + "pct_cuda_time": 0.06594044944871705, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 3.805, + "pct_cuda_time": 0.06594044944871705, + "trace": "_C::rotary_embedding(int64[1], bfloat16[1, 4096], bfloat16[1, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 164.776, + "cuda_time_us": 10.212, + "pct_cuda_time": 0.1769734217530351, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 2.603, + "pct_cuda_time": 0.045109852802893695, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[1], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_QKV_mfma4_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0, __hip_bfloat16, 16, 128, 256, false, 4>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16 const*, int, float, int const*, int const*, int const*, int, float const*, int, int, int, float*, float*, __hip_bfloat16*, __hip_bfloat16*, int, float const*, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 5.167, + "pct_cuda_time": 0.08954383766137215, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_reduce_kernel<__hip_bfloat16, __hip_bfloat16, 128, 128, 256, 1>(__hip_bfloat16*, float const*, float const*, __hip_bfloat16 const*, int const*, int const*, int, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 2.442, + "pct_cuda_time": 0.04231973128876927, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 57.596, + "cuda_time_us": 14.551, + "pct_cuda_time": 0.25216806305605305, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 14.551, + "pct_cuda_time": 0.25216806305605305, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 28.512, + "cuda_time_us": 4.566, + "pct_cuda_time": 0.07912853933846047, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.566, + "pct_cuda_time": 0.07912853933846047, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 148.161, + "cuda_time_us": 105.274, + "pct_cuda_time": 1.8243928712915216, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 57.075, + "cuda_time_us": 62.883, + "pct_cuda_time": 1.0897590756067477, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 62.883, + "pct_cuda_time": 1.0897590756067477, + "trace": "_rocm_C::wvSplitK(bfloat16[28672, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 19.93, + "cuda_time_us": 7.971, + "pct_cuda_time": 0.13813700987009822, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 7.971, + "pct_cuda_time": 0.13813700987009822, + "trace": "_C::silu_and_mul(bfloat16[1, 14336], bfloat16[1, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 49.824, + "cuda_time_us": 34.42, + "pct_cuda_time": 0.5964967858146758, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 34.42, + "pct_cuda_time": 0.5964967858146758, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 14336], bfloat16[1, 14336], 304)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 539.666, + "cuda_time_us": 157.594, + "pct_cuda_time": 2.7310957136454967, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 25.999, + "cuda_time_us": 4.486, + "pct_cuda_time": 0.07774214355504461, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.486, + "pct_cuda_time": 0.07774214355504461, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 324.464, + "cuda_time_us": 42.548, + "pct_cuda_time": 0.7373545974097275, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 51.267, + "cuda_time_us": 13.579, + "pct_cuda_time": 0.2353233542875503, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 13.579, + "pct_cuda_time": 0.2353233542875503, + "trace": "_rocm_C::wvSplitK(bfloat16[6144, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 24.647, + "cuda_time_us": 3.644, + "pct_cuda_time": 0.06315032793459263, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 3.644, + "pct_cuda_time": 0.06315032793459263, + "trace": "_C::rotary_embedding(int64[1], bfloat16[1, 4096], bfloat16[1, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 157.124, + "cuda_time_us": 10.693999999999999, + "pct_cuda_time": 0.18532645634811568, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 2.844, + "pct_cuda_time": 0.049286370100433984, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[1], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_QKV_mfma4_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0, __hip_bfloat16, 16, 128, 256, false, 4>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16 const*, int, float, int const*, int const*, int const*, int, float const*, int, int, int, float*, float*, __hip_bfloat16*, __hip_bfloat16*, int, float const*, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 5.207, + "pct_cuda_time": 0.09023703555308007, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_reduce_kernel<__hip_bfloat16, __hip_bfloat16, 128, 128, 256, 1>(__hip_bfloat16*, float const*, float const*, __hip_bfloat16 const*, int const*, int const*, int, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 2.643, + "pct_cuda_time": 0.04580305069460162, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 57.936, + "cuda_time_us": 14.631, + "pct_cuda_time": 0.2535544588394689, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 14.631, + "pct_cuda_time": 0.2535544588394689, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 29.764, + "cuda_time_us": 4.566, + "pct_cuda_time": 0.07912853933846047, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.566, + "pct_cuda_time": 0.07912853933846047, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 143.143, + "cuda_time_us": 105.994, + "pct_cuda_time": 1.8368704333422643, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 56.204, + "cuda_time_us": 62.121, + "pct_cuda_time": 1.0765536557697115, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 62.121, + "pct_cuda_time": 1.0765536557697115, + "trace": "_rocm_C::wvSplitK(bfloat16[28672, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 20.09, + "cuda_time_us": 8.091, + "pct_cuda_time": 0.14021660354522197, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 8.091, + "pct_cuda_time": 0.14021660354522197, + "trace": "_C::silu_and_mul(bfloat16[1, 14336], bfloat16[1, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 50.726, + "cuda_time_us": 35.782, + "pct_cuda_time": 0.6201001740273308, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 35.782, + "pct_cuda_time": 0.6201001740273308, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 14336], bfloat16[1, 14336], 304)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 554.738, + "cuda_time_us": 162.251, + "pct_cuda_time": 2.811801278187593, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 28.462, + "cuda_time_us": 4.486, + "pct_cuda_time": 0.07774214355504461, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.486, + "pct_cuda_time": 0.07774214355504461, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 330.854, + "cuda_time_us": 46.092, + "pct_cuda_time": 0.7987719306150504, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 50.326, + "cuda_time_us": 13.98, + "pct_cuda_time": 0.24227266315192234, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 13.98, + "pct_cuda_time": 0.24227266315192234, + "trace": "_rocm_C::wvSplitK(bfloat16[6144, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 24.667, + "cuda_time_us": 3.965, + "pct_cuda_time": 0.06871324101554878, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 3.965, + "pct_cuda_time": 0.06871324101554878, + "trace": "_C::rotary_embedding(int64[1], bfloat16[1, 4096], bfloat16[1, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 161.842, + "cuda_time_us": 13.737000000000002, + "pct_cuda_time": 0.23806148595979668, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 2.883, + "pct_cuda_time": 0.04996223804484922, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[1], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_QKV_mfma4_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0, __hip_bfloat16, 16, 128, 256, false, 4>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16 const*, int, float, int const*, int const*, int const*, int, float const*, int, int, int, float*, float*, __hip_bfloat16*, __hip_bfloat16*, int, float const*, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 6.929, + "pct_cuda_time": 0.12007920479110658, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_reduce_kernel<__hip_bfloat16, __hip_bfloat16, 128, 128, 256, 1>(__hip_bfloat16*, float const*, float const*, __hip_bfloat16 const*, int const*, int const*, int, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 3.925, + "pct_cuda_time": 0.06802004312384086, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 62.534, + "cuda_time_us": 14.41, + "pct_cuda_time": 0.24972454048778261, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 14.41, + "pct_cuda_time": 0.24972454048778261, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 29.283, + "cuda_time_us": 4.606, + "pct_cuda_time": 0.0798217372301684, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.606, + "pct_cuda_time": 0.0798217372301684, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 146.289, + "cuda_time_us": 107.06700000000001, + "pct_cuda_time": 1.8554654667873296, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 56.754, + "cuda_time_us": 63.214, + "pct_cuda_time": 1.0954952881606308, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 63.214, + "pct_cuda_time": 1.0954952881606308, + "trace": "_rocm_C::wvSplitK(bfloat16[28672, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 20.501, + "cuda_time_us": 8.051, + "pct_cuda_time": 0.13952340565351407, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 8.051, + "pct_cuda_time": 0.13952340565351407, + "trace": "_C::silu_and_mul(bfloat16[1, 14336], bfloat16[1, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 51.177, + "cuda_time_us": 35.802, + "pct_cuda_time": 0.6204467729731847, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 35.802, + "pct_cuda_time": 0.6204467729731847, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 14336], bfloat16[1, 14336], 304)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 568.028, + "cuda_time_us": 158.034, + "pct_cuda_time": 2.738720890454284, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 28.703, + "cuda_time_us": 4.485, + "pct_cuda_time": 0.07772481360775191, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.485, + "pct_cuda_time": 0.07772481360775191, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 351.144, + "cuda_time_us": 44.93000000000001, + "pct_cuda_time": 0.778634531860935, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 51.707, + "cuda_time_us": 14.06, + "pct_cuda_time": 0.24365905893533818, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 14.06, + "pct_cuda_time": 0.24365905893533818, + "trace": "_rocm_C::wvSplitK(bfloat16[6144, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 26.099, + "cuda_time_us": 3.604, + "pct_cuda_time": 0.062457130042884695, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 3.604, + "pct_cuda_time": 0.062457130042884695, + "trace": "_C::rotary_embedding(int64[1], bfloat16[1, 4096], bfloat16[1, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 176.894, + "cuda_time_us": 13.257000000000001, + "pct_cuda_time": 0.22974311125930147, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 2.603, + "pct_cuda_time": 0.045109852802893695, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[1], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_QKV_mfma4_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0, __hip_bfloat16, 16, 128, 256, false, 4>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16 const*, int, float, int const*, int const*, int const*, int, float const*, int, int, int, float*, float*, __hip_bfloat16*, __hip_bfloat16*, int, float const*, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 6.649, + "pct_cuda_time": 0.11522681954915105, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_reduce_kernel<__hip_bfloat16, __hip_bfloat16, 128, 128, 256, 1>(__hip_bfloat16*, float const*, float const*, __hip_bfloat16 const*, int const*, int const*, int, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 4.005, + "pct_cuda_time": 0.06940643890725672, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 60.14, + "cuda_time_us": 14.009, + "pct_cuda_time": 0.24277523162341058, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 14.009, + "pct_cuda_time": 0.24277523162341058, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 27.681, + "cuda_time_us": 4.566, + "pct_cuda_time": 0.07912853933846047, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.566, + "pct_cuda_time": 0.07912853933846047, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 143.654, + "cuda_time_us": 104.053, + "pct_cuda_time": 1.8032330056471368, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 52.989, + "cuda_time_us": 60.98, + "pct_cuda_time": 1.0567801859087427, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 60.98, + "pct_cuda_time": 1.0567801859087427, + "trace": "_rocm_C::wvSplitK(bfloat16[28672, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 20.561, + "cuda_time_us": 8.051, + "pct_cuda_time": 0.13952340565351407, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 8.051, + "pct_cuda_time": 0.13952340565351407, + "trace": "_C::silu_and_mul(bfloat16[1, 14336], bfloat16[1, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 50.095, + "cuda_time_us": 35.022, + "pct_cuda_time": 0.6069294140848801, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 35.022, + "pct_cuda_time": 0.6069294140848801, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 14336], bfloat16[1, 14336], 304)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "LlamaDecoderLayer", + "cpu_time_us": 636.3, + "cuda_time_us": 160.00900000000001, + "pct_cuda_time": 2.7729475363573637, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 27.211, + "cuda_time_us": 4.365, + "pct_cuda_time": 0.07564521993262811, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.365, + "pct_cuda_time": 0.07564521993262811, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaAttention", + "cpu_time_us": 416.822, + "cuda_time_us": 46.814, + "pct_cuda_time": 0.8112841525603786, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "QKVParallelLinear(weight=bfloat16[6144, 4096])", + "cpu_time_us": 49.103, + "cuda_time_us": 15.462, + "pct_cuda_time": 0.2679556450397012, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 15.462, + "pct_cuda_time": 0.2679556450397012, + "trace": "_rocm_C::wvSplitK(bfloat16[6144, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Llama3RotaryEmbedding", + "cpu_time_us": 25.568, + "cuda_time_us": 3.925, + "pct_cuda_time": 0.06802004312384086, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::rotary_embedding_kernel(long const*, c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, int, long, long, int, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 3.925, + "pct_cuda_time": 0.06802004312384086, + "trace": "_C::rotary_embedding(int64[1], bfloat16[1, 4096], bfloat16[1, 1024], 128, bfloat16[131072, 128], True)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Attention", + "cpu_time_us": 247.229, + "cuda_time_us": 13.417, + "pct_cuda_time": 0.2325159028261332, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void vllm::reshape_and_cache_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, __hip_bfloat16*, long const*, int, int, int, int, int, int, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 2.643, + "pct_cuda_time": 0.04580305069460162, + "trace": "_C_cache_ops::reshape_and_cache(bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], int64[1], None, float32[], float32[]) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_QKV_mfma4_kernel<__hip_bfloat16, __hip_bfloat16, (vllm::Fp8KVCacheDataType)0, __hip_bfloat16, 16, 128, 256, false, 4>(__hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16 const*, int, float, int const*, int const*, int const*, int, float const*, int, int, int, float*, float*, __hip_bfloat16*, __hip_bfloat16*, int, float const*, float const*, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 6.89, + "pct_cuda_time": 0.11940333684669131, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void paged_attention_ll4mi_reduce_kernel<__hip_bfloat16, __hip_bfloat16, 128, 128, 256, 1>(__hip_bfloat16*, float const*, float const*, __hip_bfloat16 const*, int const*, int const*, int, float const*)", + "cpu_time_us": 0, + "cuda_time_us": 3.884, + "pct_cuda_time": 0.06730951528484022, + "trace": "_rocm_C::paged_attention(bfloat16[1, 32, 128], float32[1, 32, 1], float32[1, 32, 1], bfloat16[1, 32, 1, 128], bfloat16[1, 32, 128], bfloat16[79739, 8, 16, 16, 8], bfloat16[79739, 8, 128, 16], 8, 0.08838834764831845, int32[1, 9], int32[1], None, 16, 129, None, None, float32[], float32[], None) <- vllm::unified_attention_with_output(bfloat16[1, 32, 128], bfloat16[1, 8, 128], bfloat16[1, 8, 128], bfloat16[1, 32, 128], None, None)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 4096])", + "cpu_time_us": 58.397, + "cuda_time_us": 14.01, + "pct_cuda_time": 0.24279256157070328, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 14.01, + "pct_cuda_time": 0.24279256157070328, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 26.199, + "cuda_time_us": 4.566, + "pct_cuda_time": 0.07912853933846047, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.566, + "pct_cuda_time": 0.07912853933846047, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "LlamaMLP", + "cpu_time_us": 148.602, + "cuda_time_us": 104.26400000000001, + "pct_cuda_time": 1.8068896245258963, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "MergedColumnParallelLinear(weight=bfloat16[28672, 4096])", + "cpu_time_us": 56.615, + "cuda_time_us": 62.453, + "pct_cuda_time": 1.0823071982708874, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 62.453, + "pct_cuda_time": 1.0823071982708874, + "trace": "_rocm_C::wvSplitK(bfloat16[28672, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "SiluAndMul", + "cpu_time_us": 20.421, + "cuda_time_us": 8.051, + "pct_cuda_time": 0.13952340565351407, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "_ZN4vllm18act_and_mul_kernelIN3c108BFloat16ETnPFT_RKS3_EXadL_ZNS_11silu_kernelIS2_EES3_S5_EELb1EEEvPS3_PS4_i", + "cpu_time_us": 0, + "cuda_time_us": 8.051, + "pct_cuda_time": 0.13952340565351407, + "trace": "_C::silu_and_mul(bfloat16[1, 14336], bfloat16[1, 28672])" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "RowParallelLinear(weight=bfloat16[4096, 14336])", + "cpu_time_us": 54.592, + "cuda_time_us": 33.76, + "pct_cuda_time": 0.5850590206014947, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 33.76, + "pct_cuda_time": 0.5850590206014947, + "trace": "_rocm_C::wvSplitK(bfloat16[4096, 14336], bfloat16[1, 14336], 304)" + }, + "children": [] + } + ] + } + ] + } + ] + }, + { + "entry": { + "name": "RMSNorm(weight=bfloat16[4096])", + "cpu_time_us": 27.781, + "cuda_time_us": 4.446, + "pct_cuda_time": 0.07704894566333666, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "std::enable_if<(((8)>(0)))&&_typeConvert::exists, void>::type vllm::fused_add_rms_norm_kernel(c10::BFloat16*, c10::BFloat16*, c10::BFloat16 const*, float, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 4.446, + "pct_cuda_time": 0.07704894566333666, + "trace": "_C::fused_add_rms_norm(bfloat16[1, 4096], bfloat16[1, 4096], bfloat16[4096], 1e-05)" + }, + "children": [] + } + ] + } + ] + }, + { + "entry": { + "name": "LogitsProcessor", + "cpu_time_us": 155.172, + "cuda_time_us": 264.814, + "pct_cuda_time": 4.58921266236861, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "void at::native::(anonymous namespace)::indexSelectSmallIndex(at::cuda::detail::TensorInfo, at::cuda::detail::TensorInfo, at::cuda::detail::TensorInfo, int, int, unsigned int, long)", + "cpu_time_us": 0, + "cuda_time_us": 2.603, + "pct_cuda_time": 0.045109852802893695, + "trace": "index_select(bfloat16[1, 4096], 0, int64[1])" + }, + "children": [] + }, + { + "entry": { + "name": "void wvSplitK_hf_sml_<__hip_bfloat16, 64, 2, 16, 8, 2, 1>(int, int, __hip_bfloat16 const*, __hip_bfloat16 const*, __hip_bfloat16*, int, int)", + "cpu_time_us": 0, + "cuda_time_us": 262.211, + "pct_cuda_time": 4.544102809565715, + "trace": "_rocm_C::wvSplitK(bfloat16[128256, 4096], bfloat16[1, 4096], 304)" + }, + "children": [] + } + ] + }, + { + "entry": { + "name": "Sampler", + "cpu_time_us": 1086.302, + "cuda_time_us": 181.41700000000003, + "pct_cuda_time": 3.143947047999449, + "trace": "" + }, + "children": [ + { + "entry": { + "name": "Memcpy DtoD (Device -> Device)", + "cpu_time_us": 0, + "cuda_time_us": 2.403, + "pct_cuda_time": 0.04164386334435403, + "trace": "copy_(bfloat16[1], bfloat16[1], True) <- _to_copy(bfloat16[1], 15, 0, None, None, True, None) <- to(bfloat16[1], 15, 0, None, None, True, False, None)" + }, + "children": [] + }, + { + "entry": { + "name": "Memcpy DtoD (Device -> Device)", + "cpu_time_us": 0, + "cuda_time_us": 2.363, + "pct_cuda_time": 0.0409506654526461, + "trace": "copy_(bfloat16[1], bfloat16[1], True) <- _to_copy(bfloat16[1], 15, 0, None, None, True, None) <- to(bfloat16[1], 15, 0, None, None, True, False, None)" + }, + "children": [] + }, + { + "entry": { + "name": "Memcpy DtoD (Device -> Device)", + "cpu_time_us": 0, + "cuda_time_us": 2.603, + "pct_cuda_time": 0.045109852802893695, + "trace": "copy_(int32[1], int32[1], True) <- _to_copy(int32[1], 3, 0, None, None, True, None) <- to(int32[1], 3, 0, None, None, True, False, None)" + }, + "children": [] + }, + { + "entry": { + "name": "Memcpy DtoD (Device -> Device)", + "cpu_time_us": 0, + "cuda_time_us": 2.362, + "pct_cuda_time": 0.0409333355053534, + "trace": "copy_(bfloat16[1], bfloat16[1], True) <- _to_copy(bfloat16[1], 15, 0, None, None, True, None) <- to(bfloat16[1], 15, 0, None, None, True, False, None)" + }, + "children": [] + }, + { + "entry": { + "name": "Memcpy DtoD (Device -> Device)", + "cpu_time_us": 0, + "cuda_time_us": 2.282, + "pct_cuda_time": 0.039546939721937534, + "trace": "copy_(bfloat16[1], bfloat16[1], True) <- _to_copy(bfloat16[1], 15, 0, None, None, True, None) <- to(bfloat16[1], 15, 0, None, None, True, False, None)" + }, + "children": [] + }, + { + "entry": { + "name": "Memcpy DtoD (Device -> Device)", + "cpu_time_us": 0, + "cuda_time_us": 2.323, + "pct_cuda_time": 0.04025746756093816, + "trace": "copy_(bfloat16[1], bfloat16[1], True) <- _to_copy(bfloat16[1], 15, 0, None, None, True, None) <- to(bfloat16[1], 15, 0, None, None, True, False, None)" + }, + "children": [] + }, + { + "entry": { + "name": "Memcpy DtoD (Device -> Device)", + "cpu_time_us": 0, + "cuda_time_us": 2.282, + "pct_cuda_time": 0.039546939721937534, + "trace": "copy_(bfloat16[1], bfloat16[1], True) <- _to_copy(bfloat16[1], 15, 0, None, None, True, None) <- to(bfloat16[1], 15, 0, None, None, True, False, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<512, 1, at::native::gpu_kernel_impl(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#7}::operator()() const::{lambda(float)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#7}::operator()() const::{lambda(float)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 2.883, + "pct_cuda_time": 0.04996223804484922, + "trace": "copy_(float32[1, 128256], bfloat16[1, 128256], False) <- _to_copy(bfloat16[1, 128256], 6, None, None, None, False, None) <- to(bfloat16[1, 128256], 6, False, False, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl > >(at::TensorIteratorBase&, at::native::BinaryFunctor > const&)::{lambda(int)#2}>(int, at::native::gpu_kernel_impl > >(at::TensorIteratorBase&, at::native::BinaryFunctor > const&)::{lambda(int)#2})", + "cpu_time_us": 0, + "cuda_time_us": 8.612, + "pct_cuda_time": 0.14924550608471782, + "trace": "div_(float32[1, 128256], bfloat16[1, 1])" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::(anonymous namespace)::cunn_SoftMaxForward<4, float, float, float, at::native::(anonymous namespace)::SoftMaxForwardEpilogue>(float*, float const*, int)", + "cpu_time_us": 0, + "cuda_time_us": 64.015, + "pct_cuda_time": 1.1093765759420822, + "trace": "_softmax(float32[1, 128256], -1, False) <- softmax(float32[1, 128256], -1, 6)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::(anonymous namespace)::cunn_SoftMaxForward<4, float, float, float, at::native::(anonymous namespace)::LogSoftMaxForwardEpilogue>(float*, float const*, int)", + "cpu_time_us": 0, + "cuda_time_us": 35.612, + "pct_cuda_time": 0.6171540829875721, + "trace": "_log_softmax(float32[1, 128256], -1, False) <- log_softmax(float32[1, 128256], -1, 6)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::elementwise_kernel<512, 1, at::native::gpu_kernel_impl(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#4}::operator()() const::{lambda(long)#1} const&)::{lambda(int)#1}>(int, at::native::gpu_kernel_impl(at::TensorIteratorBase&, at::native::direct_copy_kernel_cuda(at::TensorIteratorBase&)::{lambda()#3}::operator()() const::{lambda()#4}::operator()() const::{lambda(long)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 2.323, + "pct_cuda_time": 0.04025746756093816, + "trace": "copy_(int64[1], int32[1], False) <- _to_copy(int32[1], 4, None, None, None, False, None) <- to(int32[1], 4, False, False, None)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::index_elementwise_kernel<128, 4, at::native::gpu_index_kernel >(at::TensorIteratorBase&, c10::ArrayRef, c10::ArrayRef)::{lambda(char*, char const*, long)#1}>(at::TensorIteratorBase&, c10::ArrayRef, c10::ArrayRef, at::native::index_kernel_impl >(at::TensorIteratorBase&, c10::ArrayRef, c10::ArrayRef)::{lambda(char*, char const*, long)#1} const&)::{lambda(int)#1}>(long, at::native::gpu_index_kernel >(at::TensorIteratorBase&, c10::ArrayRef, c10::ArrayRef)::{lambda(char*, char const*, long)#1}>(at::TensorIteratorBase&, c10::ArrayRef, c10::ArrayRef, at::native::index_kernel_impl >(at::TensorIteratorBase&, c10::ArrayRef, c10::ArrayRef)::{lambda(char*, char const*, long)#1} const&)::{lambda(int)#1})", + "cpu_time_us": 0, + "cuda_time_us": 5.287, + "pct_cuda_time": 0.09162343133649595, + "trace": "index(float32[1, 128256], None)" + }, + "children": [] + }, + { + "entry": { + "name": "void at::native::reduce_kernel<512, 1, at::native::ReduceOp, unsigned int, long, 4, 4> >(at::native::ReduceOp, unsigned int, long, 4, 4>)", + "cpu_time_us": 0, + "cuda_time_us": 41.341, + "pct_cuda_time": 0.7164373510274407, + "trace": "argmax(float32[1, 128256], -1, False)" + }, + "children": [] + }, + { + "entry": { + "name": "Memcpy DtoH (Device -> Host)", + "cpu_time_us": 0, + "cuda_time_us": 4.726, + "pct_cuda_time": 0.0819013309052922, + "trace": "copy_(int64[1], int64[1], False) <- _to_copy(int64[1], 4, 0, None, None, False, None) <- to(int64[1], 4, 0, None, None, False, False, None)" + }, + "children": [] + } + ] + } + ] + } +} \ No newline at end of file