idx
int64 0
2.11M
| name
stringlengths 1
118k
| code
stringlengths 6
516k
| asm
stringlengths 21
4.64M
| file
stringlengths 39
143
| opt
stringclasses 1
value | path
stringlengths 20
133
|
|---|---|---|---|---|---|---|
2,113,200 |
ncnn::UnaryOp_x86_functor::unary_op_neg::func_pack4(float vector[4] const&) const
|
__m128 func_pack4(const __m128& x) const
{
return _mm_sub_ps(_mm_setzero_ps(), x);
}
|
movq %rdi, -0x40(%rsp)
movq %rsi, -0x48(%rsp)
xorps %xmm0, %xmm0
movaps %xmm0, -0x18(%rsp)
movaps -0x18(%rsp), %xmm1
movq -0x48(%rsp), %rax
movaps (%rax), %xmm0
movaps %xmm1, -0x28(%rsp)
movaps %xmm0, -0x38(%rsp)
movaps -0x28(%rsp), %xmm0
subps -0x38(%rsp), %xmm0
retq
nopw %cs:(%rax,%rax)
nop
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,201 |
ncnn::UnaryOp_x86_functor::unary_op_neg::func(float const&) const
|
float func(const float& x) const
{
return -x;
}
|
movq %rdi, -0x8(%rsp)
movq %rsi, -0x10(%rsp)
movq -0x10(%rsp), %rax
movss (%rax), %xmm0
movd %xmm0, %eax
xorl $0x80000000, %eax # imm = 0x80000000
movd %eax, %xmm0
retq
nopw %cs:(%rax,%rax)
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,202 |
ncnn::UnaryOp_x86_functor::unary_op_floor::func_pack4(float vector[4] const&) const
|
__m128 func_pack4(const __m128& x) const
{
#if __SSE4_1__
return _mm_floor_ps(x);
#endif // __SSE4_1__
// Use negative zero as the sign bit mask.
const __m128 magic_negative_zero = _mm_set_ps1(-0.0f);
// The smallest float number that have no fractional part. (2^23)
const __m128 magic_smallest_no_fraction = _mm_set_ps1(8388608.0f);
// absolute = abs(x);
__m128 absolute = _mm_andnot_ps(magic_negative_zero, x);
// negative_mask = magic_negative_zero && x;
__m128 negative_mask = _mm_and_ps(magic_negative_zero, x);
// no_fraction = (magic_smallest_no_fraction < absolute);
__m128 no_fraction = _mm_cmplt_ps(magic_smallest_no_fraction, absolute);
// truncated = static_cast<float>(static_cast<uint32_t>(absolute));
__m128 truncated = _mm_cvtepi32_ps(_mm_cvttps_epi32(absolute));
// truncated_with_sign = (truncated || negative_mask);
__m128 truncated_with_sign = _mm_or_ps(truncated, negative_mask);
// negative_fix = ((x < truncated_with_sign) ? 1.0f : 0.0f);
__m128 negative_fix = _mm_and_ps(
_mm_cmplt_ps(x, truncated_with_sign),
_mm_set_ps1(1.0f));
// fixed_result = truncated_with_sign - negative_fix;
__m128 fixed_result = _mm_sub_ps(truncated_with_sign, negative_fix);
// return ((x && no_fraction) || (!no_fraction && fixed_result));
return _mm_or_ps(
_mm_and_ps(x, no_fraction),
_mm_andnot_ps(no_fraction, fixed_result));
}
|
subq $0x1f8, %rsp # imm = 0x1F8
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
movl $0x80000000, 0x2c(%rsp) # imm = 0x80000000
movss 0x2c(%rsp), %xmm0
movss %xmm0, 0x8c(%rsp)
movss 0x8c(%rsp), %xmm0
shufps $0x0, %xmm0, %xmm0 # xmm0 = xmm0[0,0,0,0]
movaps %xmm0, 0x70(%rsp)
movaps 0x70(%rsp), %xmm0
movaps %xmm0, (%rsp)
movl $0x4b000000, 0x28(%rsp) # imm = 0x4B000000
movss 0x28(%rsp), %xmm0
movss %xmm0, 0xac(%rsp)
movss 0xac(%rsp), %xmm0
shufps $0x0, %xmm0, %xmm0 # xmm0 = xmm0[0,0,0,0]
movaps %xmm0, 0x90(%rsp)
movaps 0x90(%rsp), %xmm0
movaps %xmm0, -0x10(%rsp)
movaps (%rsp), %xmm1
movq 0x10(%rsp), %rax
movaps (%rax), %xmm0
movaps %xmm1, 0x60(%rsp)
movaps %xmm0, 0x50(%rsp)
movaps 0x60(%rsp), %xmm0
pcmpeqd %xmm2, %xmm2
movaps 0x50(%rsp), %xmm1
pandn %xmm1, %xmm0
movaps %xmm0, -0x20(%rsp)
movaps (%rsp), %xmm1
movq 0x10(%rsp), %rax
movaps (%rax), %xmm0
movaps %xmm1, 0x140(%rsp)
movaps %xmm0, 0x130(%rsp)
movaps 0x140(%rsp), %xmm0
movaps 0x130(%rsp), %xmm1
pand %xmm1, %xmm0
movaps %xmm0, -0x30(%rsp)
movaps -0x10(%rsp), %xmm1
movaps -0x20(%rsp), %xmm0
movaps %xmm1, 0x180(%rsp)
movaps %xmm0, 0x170(%rsp)
movaps 0x180(%rsp), %xmm0
movaps 0x170(%rsp), %xmm1
cmpltps %xmm1, %xmm0
movaps %xmm0, -0x40(%rsp)
movaps -0x20(%rsp), %xmm0
movaps %xmm0, 0x1a0(%rsp)
cvttps2dq 0x1a0(%rsp), %xmm0
movaps %xmm0, 0x190(%rsp)
cvtdq2ps 0x190(%rsp), %xmm0
movaps %xmm0, -0x50(%rsp)
movaps -0x50(%rsp), %xmm1
movaps -0x30(%rsp), %xmm0
movaps %xmm1, 0x1e0(%rsp)
movaps %xmm0, 0x1d0(%rsp)
movaps 0x1e0(%rsp), %xmm0
movaps 0x1d0(%rsp), %xmm1
por %xmm1, %xmm0
movaps %xmm0, -0x60(%rsp)
movq 0x10(%rsp), %rax
movaps (%rax), %xmm1
movaps -0x60(%rsp), %xmm0
movaps %xmm1, 0x160(%rsp)
movaps %xmm0, 0x150(%rsp)
movaps 0x160(%rsp), %xmm1
movaps 0x150(%rsp), %xmm0
cmpltps %xmm0, %xmm1
movl $0x3f800000, 0x24(%rsp) # imm = 0x3F800000
movss 0x24(%rsp), %xmm0
movss %xmm0, 0xcc(%rsp)
movss 0xcc(%rsp), %xmm0
shufps $0x0, %xmm0, %xmm0 # xmm0 = xmm0[0,0,0,0]
movaps %xmm0, 0xb0(%rsp)
movaps 0xb0(%rsp), %xmm0
movaps %xmm1, 0x120(%rsp)
movaps %xmm0, 0x110(%rsp)
movaps 0x120(%rsp), %xmm0
movaps 0x110(%rsp), %xmm1
pand %xmm1, %xmm0
movaps %xmm0, -0x70(%rsp)
movaps -0x60(%rsp), %xmm1
movaps -0x70(%rsp), %xmm0
movaps %xmm1, 0xe0(%rsp)
movaps %xmm0, 0xd0(%rsp)
movaps 0xe0(%rsp), %xmm0
movaps 0xd0(%rsp), %xmm1
subps %xmm1, %xmm0
movaps %xmm0, -0x80(%rsp)
movq 0x10(%rsp), %rax
movaps (%rax), %xmm1
movaps -0x40(%rsp), %xmm0
movaps %xmm1, 0x100(%rsp)
movaps %xmm0, 0xf0(%rsp)
movaps 0x100(%rsp), %xmm1
movaps 0xf0(%rsp), %xmm0
pand %xmm0, %xmm1
movaps -0x40(%rsp), %xmm3
movaps -0x80(%rsp), %xmm0
movaps %xmm3, 0x40(%rsp)
movaps %xmm0, 0x30(%rsp)
movaps 0x40(%rsp), %xmm0
pxor %xmm2, %xmm0
movaps 0x30(%rsp), %xmm2
pand %xmm2, %xmm0
movaps %xmm1, 0x1c0(%rsp)
movaps %xmm0, 0x1b0(%rsp)
movaps 0x1c0(%rsp), %xmm0
movaps 0x1b0(%rsp), %xmm1
por %xmm1, %xmm0
addq $0x1f8, %rsp # imm = 0x1F8
retq
nopw %cs:(%rax,%rax)
nopl (%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,203 |
ncnn::UnaryOp_x86_functor::unary_op_floor::func(float const&) const
|
float func(const float& x) const
{
return (float)floor(x);
}
|
subq $0x18, %rsp
movq %rdi, 0x10(%rsp)
movq %rsi, 0x8(%rsp)
movq 0x8(%rsp), %rax
movss (%rax), %xmm0
callq 0x8e720
addq $0x18, %rsp
retq
nopw %cs:(%rax,%rax)
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,204 |
ncnn::UnaryOp_x86_functor::unary_op_ceil::func_pack4(float vector[4] const&) const
|
__m128 func_pack4(const __m128& x) const
{
#if __SSE4_1__
return _mm_ceil_ps(x);
#endif // __SSE4_1__
// Use negative zero as the sign bit mask.
const __m128 magic_negative_zero = _mm_set_ps1(-0.0f);
// The smallest float number that have no fractional part. (2^23)
const __m128 magic_smallest_no_fraction = _mm_set_ps1(8388608.0f);
// absolute = abs(x);
__m128 absolute = _mm_andnot_ps(magic_negative_zero, x);
// negative_mask = magic_negative_zero && x;
__m128 negative_mask = _mm_and_ps(magic_negative_zero, x);
// no_fraction = (magic_smallest_no_fraction < absolute);
__m128 no_fraction = _mm_cmplt_ps(magic_smallest_no_fraction, absolute);
// truncated = static_cast<float>(static_cast<uint32_t>(absolute));
__m128 truncated = _mm_cvtepi32_ps(_mm_cvttps_epi32(absolute));
// truncated_with_sign = (truncated || negative_mask);
__m128 truncated_with_sign = _mm_or_ps(truncated, negative_mask);
// positive_fix = ((x > -0.0f) && (x > truncated_with_sign) ? -1.0f : 0.0f);
__m128 positive_fix = _mm_and_ps(
_mm_and_ps(
_mm_cmpgt_ps(x, magic_negative_zero),
_mm_cmpgt_ps(x, truncated_with_sign)),
_mm_set_ps1(-1.0f));
// fixed_result = truncated_with_sign - positive_fix;
__m128 fixed_result = _mm_sub_ps(truncated_with_sign, positive_fix);
// return ((x && no_fraction) || (!no_fraction && fixed_result));
return _mm_or_ps(
_mm_and_ps(x, no_fraction),
_mm_andnot_ps(no_fraction, fixed_result));
}
|
subq $0x238, %rsp # imm = 0x238
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
movl $0x80000000, 0x2c(%rsp) # imm = 0x80000000
movss 0x2c(%rsp), %xmm0
movss %xmm0, 0x8c(%rsp)
movss 0x8c(%rsp), %xmm0
shufps $0x0, %xmm0, %xmm0 # xmm0 = xmm0[0,0,0,0]
movaps %xmm0, 0x70(%rsp)
movaps 0x70(%rsp), %xmm0
movaps %xmm0, (%rsp)
movl $0x4b000000, 0x28(%rsp) # imm = 0x4B000000
movss 0x28(%rsp), %xmm0
movss %xmm0, 0xac(%rsp)
movss 0xac(%rsp), %xmm0
shufps $0x0, %xmm0, %xmm0 # xmm0 = xmm0[0,0,0,0]
movaps %xmm0, 0x90(%rsp)
movaps 0x90(%rsp), %xmm0
movaps %xmm0, -0x10(%rsp)
movaps (%rsp), %xmm1
movq 0x10(%rsp), %rax
movaps (%rax), %xmm0
movaps %xmm1, 0x60(%rsp)
movaps %xmm0, 0x50(%rsp)
movaps 0x60(%rsp), %xmm0
pcmpeqd %xmm2, %xmm2
movaps 0x50(%rsp), %xmm1
pandn %xmm1, %xmm0
movaps %xmm0, -0x20(%rsp)
movaps (%rsp), %xmm1
movq 0x10(%rsp), %rax
movaps (%rax), %xmm0
movaps %xmm1, 0x160(%rsp)
movaps %xmm0, 0x150(%rsp)
movaps 0x160(%rsp), %xmm0
movaps 0x150(%rsp), %xmm1
pand %xmm1, %xmm0
movaps %xmm0, -0x30(%rsp)
movaps -0x10(%rsp), %xmm1
movaps -0x20(%rsp), %xmm0
movaps %xmm1, 0x180(%rsp)
movaps %xmm0, 0x170(%rsp)
movaps 0x180(%rsp), %xmm0
movaps 0x170(%rsp), %xmm1
cmpltps %xmm1, %xmm0
movaps %xmm0, -0x40(%rsp)
movaps -0x20(%rsp), %xmm0
movaps %xmm0, 0x1a0(%rsp)
cvttps2dq 0x1a0(%rsp), %xmm0
movaps %xmm0, 0x190(%rsp)
cvtdq2ps 0x190(%rsp), %xmm0
movaps %xmm0, -0x50(%rsp)
movaps -0x50(%rsp), %xmm1
movaps -0x30(%rsp), %xmm0
movaps %xmm1, 0x1e0(%rsp)
movaps %xmm0, 0x1d0(%rsp)
movaps 0x1e0(%rsp), %xmm0
movaps 0x1d0(%rsp), %xmm1
por %xmm1, %xmm0
movaps %xmm0, -0x60(%rsp)
movq 0x10(%rsp), %rax
movaps (%rax), %xmm1
movaps (%rsp), %xmm0
movaps %xmm1, 0x220(%rsp)
movaps %xmm0, 0x210(%rsp)
movaps 0x210(%rsp), %xmm1
movaps 0x220(%rsp), %xmm0
cmpltps %xmm0, %xmm1
movq 0x10(%rsp), %rax
movaps (%rax), %xmm3
movaps -0x60(%rsp), %xmm0
movaps %xmm3, 0x200(%rsp)
movaps %xmm0, 0x1f0(%rsp)
movaps 0x1f0(%rsp), %xmm0
movaps 0x200(%rsp), %xmm3
cmpltps %xmm3, %xmm0
movaps %xmm1, 0x140(%rsp)
movaps %xmm0, 0x130(%rsp)
movaps 0x140(%rsp), %xmm1
movaps 0x130(%rsp), %xmm0
pand %xmm0, %xmm1
movl $0xbf800000, 0x24(%rsp) # imm = 0xBF800000
movss 0x24(%rsp), %xmm0
movss %xmm0, 0xcc(%rsp)
movss 0xcc(%rsp), %xmm0
shufps $0x0, %xmm0, %xmm0 # xmm0 = xmm0[0,0,0,0]
movaps %xmm0, 0xb0(%rsp)
movaps 0xb0(%rsp), %xmm0
movaps %xmm1, 0x120(%rsp)
movaps %xmm0, 0x110(%rsp)
movaps 0x120(%rsp), %xmm0
movaps 0x110(%rsp), %xmm1
pand %xmm1, %xmm0
movaps %xmm0, -0x70(%rsp)
movaps -0x60(%rsp), %xmm1
movaps -0x70(%rsp), %xmm0
movaps %xmm1, 0xe0(%rsp)
movaps %xmm0, 0xd0(%rsp)
movaps 0xe0(%rsp), %xmm0
movaps 0xd0(%rsp), %xmm1
subps %xmm1, %xmm0
movaps %xmm0, -0x80(%rsp)
movq 0x10(%rsp), %rax
movaps (%rax), %xmm1
movaps -0x40(%rsp), %xmm0
movaps %xmm1, 0x100(%rsp)
movaps %xmm0, 0xf0(%rsp)
movaps 0x100(%rsp), %xmm1
movaps 0xf0(%rsp), %xmm0
pand %xmm0, %xmm1
movaps -0x40(%rsp), %xmm3
movaps -0x80(%rsp), %xmm0
movaps %xmm3, 0x40(%rsp)
movaps %xmm0, 0x30(%rsp)
movaps 0x40(%rsp), %xmm0
pxor %xmm2, %xmm0
movaps 0x30(%rsp), %xmm2
pand %xmm2, %xmm0
movaps %xmm1, 0x1c0(%rsp)
movaps %xmm0, 0x1b0(%rsp)
movaps 0x1c0(%rsp), %xmm0
movaps 0x1b0(%rsp), %xmm1
por %xmm1, %xmm0
addq $0x238, %rsp # imm = 0x238
retq
nopw %cs:(%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,205 |
ncnn::UnaryOp_x86_functor::unary_op_ceil::func(float const&) const
|
float func(const float& x) const
{
return (float)ceil(x);
}
|
subq $0x18, %rsp
movq %rdi, 0x10(%rsp)
movq %rsi, 0x8(%rsp)
movq 0x8(%rsp), %rax
movss (%rax), %xmm0
callq 0x1058980
addq $0x18, %rsp
retq
nopw %cs:(%rax,%rax)
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,206 |
ncnn::UnaryOp_x86_functor::unary_op_square::func_pack4(float vector[4] const&) const
|
__m128 func_pack4(const __m128& x) const
{
return _mm_mul_ps(x, x);
}
|
movq %rdi, -0x30(%rsp)
movq %rsi, -0x38(%rsp)
movq -0x38(%rsp), %rax
movaps (%rax), %xmm1
movq -0x38(%rsp), %rax
movaps (%rax), %xmm0
movaps %xmm1, -0x18(%rsp)
movaps %xmm0, -0x28(%rsp)
movaps -0x18(%rsp), %xmm0
mulps -0x28(%rsp), %xmm0
retq
nop
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,207 |
ncnn::UnaryOp_x86_functor::unary_op_square::func(float const&) const
|
float func(const float& x) const
{
return x * x;
}
|
movq %rdi, -0x8(%rsp)
movq %rsi, -0x10(%rsp)
movq -0x10(%rsp), %rax
movss (%rax), %xmm0
movq -0x10(%rsp), %rax
mulss (%rax), %xmm0
retq
nopl (%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,208 |
ncnn::UnaryOp_x86_functor::unary_op_sqrt::func_pack4(float vector[4] const&) const
|
__m128 func_pack4(const __m128& x) const
{
return _mm_sqrt_ps(x);
}
|
movq %rdi, -0x20(%rsp)
movq %rsi, -0x28(%rsp)
movq -0x28(%rsp), %rax
movaps (%rax), %xmm0
movaps %xmm0, -0x18(%rsp)
movaps -0x18(%rsp), %xmm2
rsqrtps %xmm2, %xmm3
movaps %xmm2, %xmm0
mulps %xmm3, %xmm0
movaps 0x7db5e4(%rip), %xmm4 # 0x1e1c9d0
movaps %xmm0, %xmm1
mulps %xmm4, %xmm1
mulps %xmm3, %xmm0
movaps 0x7db5e4(%rip), %xmm3 # 0x1e1c9e0
addps %xmm3, %xmm0
mulps %xmm0, %xmm1
movaps 0x7d6d17(%rip), %xmm0 # 0x1e18120
pand %xmm0, %xmm2
movaps 0x7ca6fc(%rip), %xmm0 # 0x1e0bb10
cmpleps %xmm2, %xmm0
pand %xmm1, %xmm0
retq
nopl (%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,209 |
ncnn::UnaryOp_x86_functor::unary_op_sqrt::func(float const&) const
|
float func(const float& x) const
{
return (float)sqrt(x);
}
|
subq $0x18, %rsp
movq %rdi, 0x10(%rsp)
movq %rsi, 0x8(%rsp)
movq 0x8(%rsp), %rax
movss (%rax), %xmm0
callq 0x671b0
addq $0x18, %rsp
retq
nopw %cs:(%rax,%rax)
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,210 |
ncnn::UnaryOp_x86_functor::unary_op_rsqrt::func_pack4(float vector[4] const&) const
|
__m128 func_pack4(const __m128& x) const
{
return _mm_rsqrt_ps(x);
}
|
movq %rdi, -0x20(%rsp)
movq %rsi, -0x28(%rsp)
movq -0x28(%rsp), %rax
movaps (%rax), %xmm0
movaps %xmm0, -0x18(%rsp)
movaps -0x18(%rsp), %xmm0
rsqrtps %xmm0, %xmm0
retq
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,211 |
ncnn::UnaryOp_x86_functor::unary_op_rsqrt::func(float const&) const
|
float func(const float& x) const
{
return (float)(1.f / sqrt(x));
}
|
subq $0x18, %rsp
movq %rdi, 0x10(%rsp)
movq %rsi, 0x8(%rsp)
movq 0x8(%rsp), %rax
movss (%rax), %xmm0
callq 0x671b0
movaps %xmm0, %xmm1
movss 0x7c7b6d(%rip), %xmm0 # 0x1e09004
divss %xmm1, %xmm0
addq $0x18, %rsp
retq
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,212 |
ncnn::UnaryOp_x86_functor::unary_op_exp::func_pack4(float vector[4] const&) const
|
__m128 func_pack4(const __m128& x) const
{
return exp_ps(x);
}
|
subq $0x568, %rsp # imm = 0x568
movq %rdi, -0x78(%rsp)
movq %rsi, -0x80(%rsp)
movq -0x80(%rsp), %rax
movaps (%rax), %xmm0
movaps %xmm0, (%rsp)
xorps %xmm0, %xmm0
movaps %xmm0, 0x30(%rsp)
movaps 0x30(%rsp), %xmm0
movaps %xmm0, -0x10(%rsp)
movaps 0x7ca56a(%rip), %xmm0 # 0x1e0ba40
movaps %xmm0, -0x40(%rsp)
movaps (%rsp), %xmm0
movaps %xmm0, 0x120(%rsp)
movaps 0x7ca562(%rip), %xmm0 # 0x1e0ba50
movaps %xmm0, 0x110(%rsp)
movaps 0x120(%rsp), %xmm0
movaps 0x110(%rsp), %xmm1
minps %xmm1, %xmm0
movaps %xmm0, (%rsp)
movaps (%rsp), %xmm0
movaps %xmm0, 0x140(%rsp)
movaps 0x7ca540(%rip), %xmm0 # 0x1e0ba60
movaps %xmm0, 0x130(%rsp)
movaps 0x140(%rsp), %xmm0
movaps 0x130(%rsp), %xmm1
maxps %xmm1, %xmm0
movaps %xmm0, (%rsp)
movaps (%rsp), %xmm0
movaps %xmm0, 0x100(%rsp)
movaps 0x7ca51e(%rip), %xmm0 # 0x1e0ba70
movaps %xmm0, 0xf0(%rsp)
movaps 0x100(%rsp), %xmm0
movaps 0xf0(%rsp), %xmm1
mulps %xmm1, %xmm0
movaps %xmm0, -0x20(%rsp)
movaps -0x20(%rsp), %xmm0
movaps %xmm0, 0x180(%rsp)
movaps 0x7ca4fa(%rip), %xmm0 # 0x1e0ba80
movaps %xmm0, 0x170(%rsp)
movaps 0x180(%rsp), %xmm1
movaps 0x170(%rsp), %xmm2
addps %xmm2, %xmm1
movaps %xmm1, -0x20(%rsp)
movaps -0x20(%rsp), %xmm1
movaps %xmm1, 0x80(%rsp)
cvttps2dq 0x80(%rsp), %xmm1
movaps %xmm1, -0x30(%rsp)
movaps -0x30(%rsp), %xmm1
movaps %xmm1, 0x60(%rsp)
cvtdq2ps 0x60(%rsp), %xmm1
movaps %xmm1, -0x10(%rsp)
movaps -0x10(%rsp), %xmm2
movaps -0x20(%rsp), %xmm1
movaps %xmm2, 0xa0(%rsp)
movaps %xmm1, 0x90(%rsp)
movaps 0x90(%rsp), %xmm1
movaps 0xa0(%rsp), %xmm2
cmpltps %xmm2, %xmm1
movaps %xmm1, -0x50(%rsp)
movaps -0x50(%rsp), %xmm2
movaps -0x40(%rsp), %xmm1
movaps %xmm2, 0x50(%rsp)
movaps %xmm1, 0x40(%rsp)
movaps 0x50(%rsp), %xmm1
movaps 0x40(%rsp), %xmm2
pand %xmm2, %xmm1
movaps %xmm1, -0x50(%rsp)
movaps -0x10(%rsp), %xmm2
movaps -0x50(%rsp), %xmm1
movaps %xmm2, 0x20(%rsp)
movaps %xmm1, 0x10(%rsp)
movaps 0x20(%rsp), %xmm1
movaps 0x10(%rsp), %xmm2
subps %xmm2, %xmm1
movaps %xmm1, -0x20(%rsp)
movaps -0x20(%rsp), %xmm2
movaps (%rsp), %xmm1
movaps %xmm2, 0x220(%rsp)
movaps 0x7ca423(%rip), %xmm2 # 0x1e0ba90
movaps %xmm2, 0x210(%rsp)
movaps %xmm1, 0x200(%rsp)
movaps 0x200(%rsp), %xmm2
movaps 0x220(%rsp), %xmm3
movaps 0x210(%rsp), %xmm1
movaps %xmm3, 0x260(%rsp)
movaps %xmm1, 0x250(%rsp)
movaps 0x260(%rsp), %xmm1
movaps 0x250(%rsp), %xmm3
mulps %xmm3, %xmm1
movaps %xmm2, 0x240(%rsp)
movaps %xmm1, 0x230(%rsp)
movaps 0x240(%rsp), %xmm1
movaps 0x230(%rsp), %xmm2
subps %xmm2, %xmm1
movaps %xmm1, (%rsp)
movaps -0x20(%rsp), %xmm2
movaps (%rsp), %xmm1
movaps %xmm2, 0x1b0(%rsp)
movaps 0x7ca3a9(%rip), %xmm2 # 0x1e0baa0
movaps %xmm2, 0x1a0(%rsp)
movaps %xmm1, 0x190(%rsp)
movaps 0x190(%rsp), %xmm2
movaps 0x1b0(%rsp), %xmm3
movaps 0x1a0(%rsp), %xmm1
movaps %xmm3, 0x1f0(%rsp)
movaps %xmm1, 0x1e0(%rsp)
movaps 0x1f0(%rsp), %xmm1
movaps 0x1e0(%rsp), %xmm3
mulps %xmm3, %xmm1
movaps %xmm2, 0x1d0(%rsp)
movaps %xmm1, 0x1c0(%rsp)
movaps 0x1d0(%rsp), %xmm1
movaps 0x1c0(%rsp), %xmm2
subps %xmm2, %xmm1
movaps %xmm1, (%rsp)
movaps (%rsp), %xmm1
movaps %xmm1, 0xe0(%rsp)
movaps %xmm1, 0xd0(%rsp)
movaps 0xe0(%rsp), %xmm1
movaps 0xd0(%rsp), %xmm2
mulps %xmm2, %xmm1
movaps %xmm1, -0x10(%rsp)
movaps 0x7ca314(%rip), %xmm1 # 0x1e0bab0
movaps %xmm1, -0x60(%rsp)
movaps -0x60(%rsp), %xmm2
movaps (%rsp), %xmm1
movaps %xmm2, 0x4c0(%rsp)
movaps %xmm1, 0x4b0(%rsp)
movaps 0x7ca2ff(%rip), %xmm1 # 0x1e0bac0
movaps %xmm1, 0x4a0(%rsp)
movaps 0x4c0(%rsp), %xmm2
movaps 0x4b0(%rsp), %xmm1
movaps %xmm2, 0x4e0(%rsp)
movaps %xmm1, 0x4d0(%rsp)
movaps 0x4e0(%rsp), %xmm2
movaps 0x4d0(%rsp), %xmm1
mulps %xmm1, %xmm2
movaps 0x4a0(%rsp), %xmm1
movaps %xmm2, 0x500(%rsp)
movaps %xmm1, 0x4f0(%rsp)
movaps 0x500(%rsp), %xmm1
movaps 0x4f0(%rsp), %xmm2
addps %xmm2, %xmm1
movaps %xmm1, -0x60(%rsp)
movaps -0x60(%rsp), %xmm2
movaps (%rsp), %xmm1
movaps %xmm2, 0x450(%rsp)
movaps %xmm1, 0x440(%rsp)
movaps 0x7ca284(%rip), %xmm1 # 0x1e0bad0
movaps %xmm1, 0x430(%rsp)
movaps 0x450(%rsp), %xmm2
movaps 0x440(%rsp), %xmm1
movaps %xmm2, 0x470(%rsp)
movaps %xmm1, 0x460(%rsp)
movaps 0x470(%rsp), %xmm2
movaps 0x460(%rsp), %xmm1
mulps %xmm1, %xmm2
movaps 0x430(%rsp), %xmm1
movaps %xmm2, 0x490(%rsp)
movaps %xmm1, 0x480(%rsp)
movaps 0x490(%rsp), %xmm1
movaps 0x480(%rsp), %xmm2
addps %xmm2, %xmm1
movaps %xmm1, -0x60(%rsp)
movaps -0x60(%rsp), %xmm2
movaps (%rsp), %xmm1
movaps %xmm2, 0x3e0(%rsp)
movaps %xmm1, 0x3d0(%rsp)
movaps 0x7ca209(%rip), %xmm1 # 0x1e0bae0
movaps %xmm1, 0x3c0(%rsp)
movaps 0x3e0(%rsp), %xmm2
movaps 0x3d0(%rsp), %xmm1
movaps %xmm2, 0x400(%rsp)
movaps %xmm1, 0x3f0(%rsp)
movaps 0x400(%rsp), %xmm2
movaps 0x3f0(%rsp), %xmm1
mulps %xmm1, %xmm2
movaps 0x3c0(%rsp), %xmm1
movaps %xmm2, 0x420(%rsp)
movaps %xmm1, 0x410(%rsp)
movaps 0x420(%rsp), %xmm1
movaps 0x410(%rsp), %xmm2
addps %xmm2, %xmm1
movaps %xmm1, -0x60(%rsp)
movaps -0x60(%rsp), %xmm2
movaps (%rsp), %xmm1
movaps %xmm2, 0x370(%rsp)
movaps %xmm1, 0x360(%rsp)
movaps 0x7ca18e(%rip), %xmm1 # 0x1e0baf0
movaps %xmm1, 0x350(%rsp)
movaps 0x370(%rsp), %xmm2
movaps 0x360(%rsp), %xmm1
movaps %xmm2, 0x390(%rsp)
movaps %xmm1, 0x380(%rsp)
movaps 0x390(%rsp), %xmm2
movaps 0x380(%rsp), %xmm1
mulps %xmm1, %xmm2
movaps 0x350(%rsp), %xmm1
movaps %xmm2, 0x3b0(%rsp)
movaps %xmm1, 0x3a0(%rsp)
movaps 0x3b0(%rsp), %xmm1
movaps 0x3a0(%rsp), %xmm2
addps %xmm2, %xmm1
movaps %xmm1, -0x60(%rsp)
movaps -0x60(%rsp), %xmm2
movaps (%rsp), %xmm1
movaps %xmm2, 0x300(%rsp)
movaps %xmm1, 0x2f0(%rsp)
movaps %xmm0, 0x2e0(%rsp)
movaps 0x300(%rsp), %xmm1
movaps 0x2f0(%rsp), %xmm0
movaps %xmm1, 0x320(%rsp)
movaps %xmm0, 0x310(%rsp)
movaps 0x320(%rsp), %xmm1
movaps 0x310(%rsp), %xmm0
mulps %xmm0, %xmm1
movaps 0x2e0(%rsp), %xmm0
movaps %xmm1, 0x340(%rsp)
movaps %xmm0, 0x330(%rsp)
movaps 0x340(%rsp), %xmm0
movaps 0x330(%rsp), %xmm1
addps %xmm1, %xmm0
movaps %xmm0, -0x60(%rsp)
movaps -0x60(%rsp), %xmm2
movaps -0x10(%rsp), %xmm1
movaps (%rsp), %xmm0
movaps %xmm2, 0x290(%rsp)
movaps %xmm1, 0x280(%rsp)
movaps %xmm0, 0x270(%rsp)
movaps 0x290(%rsp), %xmm1
movaps 0x280(%rsp), %xmm0
movaps %xmm1, 0x2b0(%rsp)
movaps %xmm0, 0x2a0(%rsp)
movaps 0x2b0(%rsp), %xmm1
movaps 0x2a0(%rsp), %xmm0
mulps %xmm0, %xmm1
movaps 0x270(%rsp), %xmm0
movaps %xmm1, 0x2d0(%rsp)
movaps %xmm0, 0x2c0(%rsp)
movaps 0x2d0(%rsp), %xmm0
movaps 0x2c0(%rsp), %xmm1
addps %xmm1, %xmm0
movaps %xmm0, -0x60(%rsp)
movaps -0x60(%rsp), %xmm1
movaps -0x40(%rsp), %xmm0
movaps %xmm1, 0x160(%rsp)
movaps %xmm0, 0x150(%rsp)
movaps 0x160(%rsp), %xmm0
movaps 0x150(%rsp), %xmm1
addps %xmm1, %xmm0
movaps %xmm0, -0x60(%rsp)
movaps -0x20(%rsp), %xmm0
movaps %xmm0, 0x70(%rsp)
cvttps2dq 0x70(%rsp), %xmm0
movaps %xmm0, -0x30(%rsp)
movaps -0x30(%rsp), %xmm0
movaps %xmm0, 0x520(%rsp)
movaps 0x7c9fcb(%rip), %xmm0 # 0x1e0bb00
movaps %xmm0, 0x510(%rsp)
movdqa 0x520(%rsp), %xmm0
movdqa 0x510(%rsp), %xmm1
paddd %xmm1, %xmm0
movdqa %xmm0, -0x30(%rsp)
movdqa -0x30(%rsp), %xmm0
movdqa %xmm0, 0x540(%rsp)
movl $0x17, 0x53c(%rsp)
movdqa 0x540(%rsp), %xmm0
movl 0x53c(%rsp), %eax
movd %eax, %xmm1
pslld %xmm1, %xmm0
movdqa %xmm0, -0x30(%rsp)
movdqa -0x30(%rsp), %xmm0
movdqa %xmm0, 0x550(%rsp)
movdqa 0x550(%rsp), %xmm0
movaps %xmm0, -0x70(%rsp)
movaps -0x60(%rsp), %xmm1
movaps -0x70(%rsp), %xmm0
movaps %xmm1, 0xc0(%rsp)
movaps %xmm0, 0xb0(%rsp)
movaps 0xc0(%rsp), %xmm0
mulps 0xb0(%rsp), %xmm0
movaps %xmm0, -0x60(%rsp)
movaps -0x60(%rsp), %xmm0
addq $0x568, %rsp # imm = 0x568
retq
nopw (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,213 |
ncnn::UnaryOp_x86_functor::unary_op_exp::func(float const&) const
|
float func(const float& x) const
{
return (float)exp(x);
}
|
subq $0x18, %rsp
movq %rdi, 0x10(%rsp)
movq %rsi, 0x8(%rsp)
movq 0x8(%rsp), %rax
movss (%rax), %xmm0
callq 0x100cf0
addq $0x18, %rsp
retq
nopw %cs:(%rax,%rax)
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,214 |
ncnn::UnaryOp_x86_functor::unary_op_log::func_pack4(float vector[4] const&) const
|
__m128 func_pack4(const __m128& x) const
{
return log_ps(x);
}
|
subq $0x768, %rsp # imm = 0x768
movq %rdi, -0x78(%rsp)
movq %rsi, -0x80(%rsp)
movq -0x80(%rsp), %rax
movaps (%rax), %xmm0
movaps %xmm0, 0x10(%rsp)
movaps 0x7c9dfb(%rip), %xmm0 # 0x1e0ba40
movaps %xmm0, -0x10(%rsp)
movaps 0x10(%rsp), %xmm1
xorps %xmm0, %xmm0
movaps %xmm0, 0x60(%rsp)
movaps 0x60(%rsp), %xmm0
movaps %xmm1, 0x700(%rsp)
movaps %xmm0, 0x6f0(%rsp)
movaps 0x700(%rsp), %xmm0
movaps 0x6f0(%rsp), %xmm1
cmpleps %xmm1, %xmm0
movaps %xmm0, -0x20(%rsp)
movaps 0x10(%rsp), %xmm0
movaps %xmm0, 0x1b0(%rsp)
movaps 0x7c9e77(%rip), %xmm0 # 0x1e0bb10
movaps %xmm0, 0x1a0(%rsp)
movaps 0x1b0(%rsp), %xmm0
movaps 0x1a0(%rsp), %xmm1
maxps %xmm1, %xmm0
movaps %xmm0, 0x10(%rsp)
movaps 0x10(%rsp), %xmm0
movaps %xmm0, 0x730(%rsp)
movaps 0x730(%rsp), %xmm0
movaps %xmm0, 0x720(%rsp)
movl $0x17, 0x71c(%rsp)
movaps 0x720(%rsp), %xmm0
movd 0x71c(%rsp), %xmm1
psrld %xmm1, %xmm0
movaps %xmm0, (%rsp)
movaps 0x10(%rsp), %xmm0
movaps %xmm0, 0xc0(%rsp)
movaps 0x7c9e12(%rip), %xmm0 # 0x1e0bb20
movaps %xmm0, 0xb0(%rsp)
movaps 0xc0(%rsp), %xmm0
movaps 0xb0(%rsp), %xmm1
pand %xmm1, %xmm0
movaps %xmm0, 0x10(%rsp)
movaps 0x10(%rsp), %xmm0
movaps %xmm0, 0x130(%rsp)
movaps 0x7c9d3d(%rip), %xmm1 # 0x1e0ba80
movaps %xmm1, 0x120(%rsp)
movaps 0x130(%rsp), %xmm0
movaps 0x120(%rsp), %xmm2
por %xmm2, %xmm0
movaps %xmm0, 0x10(%rsp)
movaps (%rsp), %xmm0
movaps %xmm0, 0x750(%rsp)
movaps 0x7c9d89(%rip), %xmm0 # 0x1e0bb00
movaps %xmm0, 0x740(%rsp)
movaps 0x750(%rsp), %xmm0
movaps 0x740(%rsp), %xmm2
psubd %xmm2, %xmm0
movaps %xmm0, (%rsp)
movaps (%rsp), %xmm0
movaps %xmm0, 0xf0(%rsp)
cvtdq2ps 0xf0(%rsp), %xmm0
movaps %xmm0, -0x30(%rsp)
movaps -0x30(%rsp), %xmm2
movaps -0x10(%rsp), %xmm0
movaps %xmm2, 0x210(%rsp)
movaps %xmm0, 0x200(%rsp)
movaps 0x210(%rsp), %xmm0
movaps 0x200(%rsp), %xmm2
addps %xmm2, %xmm0
movaps %xmm0, -0x30(%rsp)
movaps 0x10(%rsp), %xmm0
movaps %xmm0, 0xe0(%rsp)
movaps 0x7c9d3a(%rip), %xmm0 # 0x1e0bb30
movaps %xmm0, 0xd0(%rsp)
movaps 0xe0(%rsp), %xmm0
movaps 0xd0(%rsp), %xmm2
cmpltps %xmm2, %xmm0
movaps %xmm0, -0x40(%rsp)
movaps 0x10(%rsp), %xmm2
movaps -0x40(%rsp), %xmm0
movaps %xmm2, 0xa0(%rsp)
movaps %xmm0, 0x90(%rsp)
movaps 0xa0(%rsp), %xmm0
movaps 0x90(%rsp), %xmm2
pand %xmm2, %xmm0
movaps %xmm0, -0x50(%rsp)
movaps 0x10(%rsp), %xmm2
movaps -0x10(%rsp), %xmm0
movaps %xmm2, 0x50(%rsp)
movaps %xmm0, 0x40(%rsp)
movaps 0x50(%rsp), %xmm0
movaps 0x40(%rsp), %xmm2
subps %xmm2, %xmm0
movaps %xmm0, 0x10(%rsp)
movaps -0x30(%rsp), %xmm2
movaps -0x10(%rsp), %xmm3
movaps -0x40(%rsp), %xmm0
movaps %xmm3, 0x80(%rsp)
movaps %xmm0, 0x70(%rsp)
movaps 0x80(%rsp), %xmm0
movaps 0x70(%rsp), %xmm3
pand %xmm3, %xmm0
movaps %xmm2, 0x30(%rsp)
movaps %xmm0, 0x20(%rsp)
movaps 0x30(%rsp), %xmm0
movaps 0x20(%rsp), %xmm2
subps %xmm2, %xmm0
movaps %xmm0, -0x30(%rsp)
movaps 0x10(%rsp), %xmm2
movaps -0x50(%rsp), %xmm0
movaps %xmm2, 0x1f0(%rsp)
movaps %xmm0, 0x1e0(%rsp)
movaps 0x1f0(%rsp), %xmm0
movaps 0x1e0(%rsp), %xmm2
addps %xmm2, %xmm0
movaps %xmm0, 0x10(%rsp)
movaps 0x10(%rsp), %xmm0
movaps %xmm0, 0x190(%rsp)
movaps %xmm0, 0x180(%rsp)
movaps 0x190(%rsp), %xmm0
movaps 0x180(%rsp), %xmm2
mulps %xmm2, %xmm0
movaps %xmm0, -0x60(%rsp)
movaps 0x7c9c21(%rip), %xmm0 # 0x1e0bb40
movaps %xmm0, -0x70(%rsp)
movaps -0x70(%rsp), %xmm2
movaps 0x10(%rsp), %xmm0
movaps %xmm2, 0x6a0(%rsp)
movaps %xmm0, 0x690(%rsp)
movaps 0x7c9c0b(%rip), %xmm0 # 0x1e0bb50
movaps %xmm0, 0x680(%rsp)
movaps 0x6a0(%rsp), %xmm2
movaps 0x690(%rsp), %xmm0
movaps %xmm2, 0x6c0(%rsp)
movaps %xmm0, 0x6b0(%rsp)
movaps 0x6c0(%rsp), %xmm2
movaps 0x6b0(%rsp), %xmm0
mulps %xmm0, %xmm2
movaps 0x680(%rsp), %xmm0
movaps %xmm2, 0x6e0(%rsp)
movaps %xmm0, 0x6d0(%rsp)
movaps 0x6e0(%rsp), %xmm0
movaps 0x6d0(%rsp), %xmm2
addps %xmm2, %xmm0
movaps %xmm0, -0x70(%rsp)
movaps -0x70(%rsp), %xmm2
movaps 0x10(%rsp), %xmm0
movaps %xmm2, 0x630(%rsp)
movaps %xmm0, 0x620(%rsp)
movaps 0x7c9b8f(%rip), %xmm0 # 0x1e0bb60
movaps %xmm0, 0x610(%rsp)
movaps 0x630(%rsp), %xmm2
movaps 0x620(%rsp), %xmm0
movaps %xmm2, 0x650(%rsp)
movaps %xmm0, 0x640(%rsp)
movaps 0x650(%rsp), %xmm2
movaps 0x640(%rsp), %xmm0
mulps %xmm0, %xmm2
movaps 0x610(%rsp), %xmm0
movaps %xmm2, 0x670(%rsp)
movaps %xmm0, 0x660(%rsp)
movaps 0x670(%rsp), %xmm0
movaps 0x660(%rsp), %xmm2
addps %xmm2, %xmm0
movaps %xmm0, -0x70(%rsp)
movaps -0x70(%rsp), %xmm2
movaps 0x10(%rsp), %xmm0
movaps %xmm2, 0x5c0(%rsp)
movaps %xmm0, 0x5b0(%rsp)
movaps 0x7c9b13(%rip), %xmm0 # 0x1e0bb70
movaps %xmm0, 0x5a0(%rsp)
movaps 0x5c0(%rsp), %xmm2
movaps 0x5b0(%rsp), %xmm0
movaps %xmm2, 0x5e0(%rsp)
movaps %xmm0, 0x5d0(%rsp)
movaps 0x5e0(%rsp), %xmm2
movaps 0x5d0(%rsp), %xmm0
mulps %xmm0, %xmm2
movaps 0x5a0(%rsp), %xmm0
movaps %xmm2, 0x600(%rsp)
movaps %xmm0, 0x5f0(%rsp)
movaps 0x600(%rsp), %xmm0
movaps 0x5f0(%rsp), %xmm2
addps %xmm2, %xmm0
movaps %xmm0, -0x70(%rsp)
movaps -0x70(%rsp), %xmm2
movaps 0x10(%rsp), %xmm0
movaps %xmm2, 0x550(%rsp)
movaps %xmm0, 0x540(%rsp)
movaps 0x7c9a97(%rip), %xmm0 # 0x1e0bb80
movaps %xmm0, 0x530(%rsp)
movaps 0x550(%rsp), %xmm2
movaps 0x540(%rsp), %xmm0
movaps %xmm2, 0x570(%rsp)
movaps %xmm0, 0x560(%rsp)
movaps 0x570(%rsp), %xmm2
movaps 0x560(%rsp), %xmm0
mulps %xmm0, %xmm2
movaps 0x530(%rsp), %xmm0
movaps %xmm2, 0x590(%rsp)
movaps %xmm0, 0x580(%rsp)
movaps 0x590(%rsp), %xmm0
movaps 0x580(%rsp), %xmm2
addps %xmm2, %xmm0
movaps %xmm0, -0x70(%rsp)
movaps -0x70(%rsp), %xmm2
movaps 0x10(%rsp), %xmm0
movaps %xmm2, 0x4e0(%rsp)
movaps %xmm0, 0x4d0(%rsp)
movaps 0x7c9a1b(%rip), %xmm0 # 0x1e0bb90
movaps %xmm0, 0x4c0(%rsp)
movaps 0x4e0(%rsp), %xmm2
movaps 0x4d0(%rsp), %xmm0
movaps %xmm2, 0x500(%rsp)
movaps %xmm0, 0x4f0(%rsp)
movaps 0x500(%rsp), %xmm2
movaps 0x4f0(%rsp), %xmm0
mulps %xmm0, %xmm2
movaps 0x4c0(%rsp), %xmm0
movaps %xmm2, 0x520(%rsp)
movaps %xmm0, 0x510(%rsp)
movaps 0x520(%rsp), %xmm0
movaps 0x510(%rsp), %xmm2
addps %xmm2, %xmm0
movaps %xmm0, -0x70(%rsp)
movaps -0x70(%rsp), %xmm2
movaps 0x10(%rsp), %xmm0
movaps %xmm2, 0x470(%rsp)
movaps %xmm0, 0x460(%rsp)
movaps 0x7c999f(%rip), %xmm0 # 0x1e0bba0
movaps %xmm0, 0x450(%rsp)
movaps 0x470(%rsp), %xmm2
movaps 0x460(%rsp), %xmm0
movaps %xmm2, 0x490(%rsp)
movaps %xmm0, 0x480(%rsp)
movaps 0x490(%rsp), %xmm2
movaps 0x480(%rsp), %xmm0
mulps %xmm0, %xmm2
movaps 0x450(%rsp), %xmm0
movaps %xmm2, 0x4b0(%rsp)
movaps %xmm0, 0x4a0(%rsp)
movaps 0x4b0(%rsp), %xmm0
movaps 0x4a0(%rsp), %xmm2
addps %xmm2, %xmm0
movaps %xmm0, -0x70(%rsp)
movaps -0x70(%rsp), %xmm2
movaps 0x10(%rsp), %xmm0
movaps %xmm2, 0x400(%rsp)
movaps %xmm0, 0x3f0(%rsp)
movaps 0x7c9923(%rip), %xmm0 # 0x1e0bbb0
movaps %xmm0, 0x3e0(%rsp)
movaps 0x400(%rsp), %xmm2
movaps 0x3f0(%rsp), %xmm0
movaps %xmm2, 0x420(%rsp)
movaps %xmm0, 0x410(%rsp)
movaps 0x420(%rsp), %xmm2
movaps 0x410(%rsp), %xmm0
mulps %xmm0, %xmm2
movaps 0x3e0(%rsp), %xmm0
movaps %xmm2, 0x440(%rsp)
movaps %xmm0, 0x430(%rsp)
movaps 0x440(%rsp), %xmm0
movaps 0x430(%rsp), %xmm2
addps %xmm2, %xmm0
movaps %xmm0, -0x70(%rsp)
movaps -0x70(%rsp), %xmm2
movaps 0x10(%rsp), %xmm0
movaps %xmm2, 0x390(%rsp)
movaps %xmm0, 0x380(%rsp)
movaps 0x7c98a7(%rip), %xmm0 # 0x1e0bbc0
movaps %xmm0, 0x370(%rsp)
movaps 0x390(%rsp), %xmm2
movaps 0x380(%rsp), %xmm0
movaps %xmm2, 0x3b0(%rsp)
movaps %xmm0, 0x3a0(%rsp)
movaps 0x3b0(%rsp), %xmm2
movaps 0x3a0(%rsp), %xmm0
mulps %xmm0, %xmm2
movaps 0x370(%rsp), %xmm0
movaps %xmm2, 0x3d0(%rsp)
movaps %xmm0, 0x3c0(%rsp)
movaps 0x3d0(%rsp), %xmm0
movaps 0x3c0(%rsp), %xmm2
addps %xmm2, %xmm0
movaps %xmm0, -0x70(%rsp)
movaps -0x70(%rsp), %xmm2
movaps 0x10(%rsp), %xmm0
movaps %xmm2, 0x170(%rsp)
movaps %xmm0, 0x160(%rsp)
movaps 0x170(%rsp), %xmm0
movaps 0x160(%rsp), %xmm2
mulps %xmm2, %xmm0
movaps %xmm0, -0x70(%rsp)
movaps -0x70(%rsp), %xmm2
movaps -0x60(%rsp), %xmm0
movaps %xmm2, 0x150(%rsp)
movaps %xmm0, 0x140(%rsp)
movaps 0x150(%rsp), %xmm0
movaps 0x140(%rsp), %xmm2
mulps %xmm2, %xmm0
movaps %xmm0, -0x70(%rsp)
movaps -0x30(%rsp), %xmm2
movaps -0x70(%rsp), %xmm0
movaps %xmm2, 0x320(%rsp)
movaps 0x7c969f(%rip), %xmm2 # 0x1e0baa0
movaps %xmm2, 0x310(%rsp)
movaps %xmm0, 0x300(%rsp)
movaps 0x320(%rsp), %xmm2
movaps 0x310(%rsp), %xmm0
movaps %xmm2, 0x340(%rsp)
movaps %xmm0, 0x330(%rsp)
movaps 0x340(%rsp), %xmm2
movaps 0x330(%rsp), %xmm0
mulps %xmm0, %xmm2
movaps 0x300(%rsp), %xmm0
movaps %xmm2, 0x360(%rsp)
movaps %xmm0, 0x350(%rsp)
movaps 0x360(%rsp), %xmm0
movaps 0x350(%rsp), %xmm2
addps %xmm2, %xmm0
movaps %xmm0, -0x70(%rsp)
movaps -0x60(%rsp), %xmm2
movaps -0x70(%rsp), %xmm0
movaps %xmm2, 0x240(%rsp)
movaps %xmm1, 0x230(%rsp)
movaps %xmm0, 0x220(%rsp)
movaps 0x220(%rsp), %xmm1
movaps 0x240(%rsp), %xmm2
movaps 0x230(%rsp), %xmm0
movaps %xmm2, 0x280(%rsp)
movaps %xmm0, 0x270(%rsp)
movaps 0x280(%rsp), %xmm0
movaps 0x270(%rsp), %xmm2
mulps %xmm2, %xmm0
movaps %xmm1, 0x260(%rsp)
movaps %xmm0, 0x250(%rsp)
movaps 0x260(%rsp), %xmm0
movaps 0x250(%rsp), %xmm1
subps %xmm1, %xmm0
movaps %xmm0, -0x70(%rsp)
movaps 0x10(%rsp), %xmm1
movaps -0x70(%rsp), %xmm0
movaps %xmm1, 0x1d0(%rsp)
movaps %xmm0, 0x1c0(%rsp)
movaps 0x1d0(%rsp), %xmm0
movaps 0x1c0(%rsp), %xmm1
addps %xmm1, %xmm0
movaps %xmm0, 0x10(%rsp)
movaps -0x30(%rsp), %xmm1
movaps 0x10(%rsp), %xmm0
movaps %xmm1, 0x2b0(%rsp)
movaps 0x7c954c(%rip), %xmm1 # 0x1e0ba90
movaps %xmm1, 0x2a0(%rsp)
movaps %xmm0, 0x290(%rsp)
movaps 0x2b0(%rsp), %xmm1
movaps 0x2a0(%rsp), %xmm0
movaps %xmm1, 0x2d0(%rsp)
movaps %xmm0, 0x2c0(%rsp)
movaps 0x2d0(%rsp), %xmm1
mulps 0x2c0(%rsp), %xmm1
movaps 0x290(%rsp), %xmm0
movaps %xmm1, 0x2f0(%rsp)
movaps %xmm0, 0x2e0(%rsp)
movaps 0x2f0(%rsp), %xmm0
addps 0x2e0(%rsp), %xmm0
movaps %xmm0, 0x10(%rsp)
movaps 0x10(%rsp), %xmm1
movaps -0x20(%rsp), %xmm0
movaps %xmm1, 0x110(%rsp)
movaps %xmm0, 0x100(%rsp)
movaps 0x110(%rsp), %xmm0
movaps 0x100(%rsp), %xmm1
por %xmm1, %xmm0
movaps %xmm0, 0x10(%rsp)
movaps 0x10(%rsp), %xmm0
addq $0x768, %rsp # imm = 0x768
retq
nopw %cs:(%rax,%rax)
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,215 |
ncnn::UnaryOp_x86_functor::unary_op_log::func(float const&) const
|
float func(const float& x) const
{
return (float)log(x);
}
|
subq $0x18, %rsp
movq %rdi, 0x10(%rsp)
movq %rsi, 0x8(%rsp)
movq 0x8(%rsp), %rax
movss (%rax), %xmm0
callq 0x100cd0
addq $0x18, %rsp
retq
nopw %cs:(%rax,%rax)
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,216 |
ncnn::UnaryOp_x86_functor::unary_op_sin::func_pack4(float vector[4] const&) const
|
__m128 func_pack4(const __m128& x) const
{
return sin_ps(x);
}
|
subq $0x708, %rsp # imm = 0x708
movq %rdi, -0x78(%rsp)
movq %rsi, -0x80(%rsp)
movq -0x80(%rsp), %rax
movaps (%rax), %xmm0
movaps %xmm0, 0x40(%rsp)
xorps %xmm0, %xmm0
movaps %xmm0, 0x70(%rsp)
movaps 0x70(%rsp), %xmm0
movaps %xmm0, 0x20(%rsp)
movaps 0x40(%rsp), %xmm0
movaps %xmm0, (%rsp)
movaps 0x40(%rsp), %xmm0
movaps %xmm0, 0xd0(%rsp)
movaps 0x7d5aa3(%rip), %xmm0 # 0x1e18120
movaps %xmm0, 0xc0(%rsp)
movaps 0xd0(%rsp), %xmm0
movaps 0xc0(%rsp), %xmm1
pand %xmm1, %xmm0
movaps %xmm0, 0x40(%rsp)
movaps (%rsp), %xmm0
movaps %xmm0, 0xb0(%rsp)
movaps 0x7cd4af(%rip), %xmm0 # 0x1e0fb60
movaps %xmm0, 0xa0(%rsp)
movaps 0xb0(%rsp), %xmm0
movaps 0xa0(%rsp), %xmm1
pand %xmm1, %xmm0
movaps %xmm0, (%rsp)
movaps 0x40(%rsp), %xmm0
movaps %xmm0, 0x190(%rsp)
movaps 0x7da30b(%rip), %xmm0 # 0x1e1c9f0
movaps %xmm0, 0x180(%rsp)
movaps 0x190(%rsp), %xmm0
movaps 0x180(%rsp), %xmm1
mulps %xmm1, %xmm0
movaps %xmm0, -0x10(%rsp)
movaps -0x10(%rsp), %xmm0
movaps %xmm0, 0xf0(%rsp)
cvttps2dq 0xf0(%rsp), %xmm0
movaps %xmm0, -0x30(%rsp)
movaps -0x30(%rsp), %xmm0
movaps %xmm0, 0x5e0(%rsp)
movaps 0x7da2cc(%rip), %xmm0 # 0x1e1ca00
movaps %xmm0, 0x5d0(%rsp)
movaps 0x5e0(%rsp), %xmm0
movaps 0x5d0(%rsp), %xmm1
paddd %xmm1, %xmm0
movaps %xmm0, -0x30(%rsp)
movaps -0x30(%rsp), %xmm0
movaps %xmm0, 0x680(%rsp)
movaps 0x7da2a7(%rip), %xmm0 # 0x1e1ca10
movaps %xmm0, 0x670(%rsp)
movaps 0x680(%rsp), %xmm0
movaps 0x670(%rsp), %xmm1
pand %xmm1, %xmm0
movaps %xmm0, -0x30(%rsp)
movaps -0x30(%rsp), %xmm0
movaps %xmm0, 0xe0(%rsp)
cvtdq2ps 0xe0(%rsp), %xmm0
movaps %xmm0, -0x10(%rsp)
movaps -0x30(%rsp), %xmm0
movaps %xmm0, 0x660(%rsp)
movaps 0x7da268(%rip), %xmm0 # 0x1e1ca20
movaps %xmm0, 0x650(%rsp)
movaps 0x660(%rsp), %xmm0
movaps 0x650(%rsp), %xmm1
pand %xmm1, %xmm0
movaps %xmm0, -0x20(%rsp)
movaps -0x20(%rsp), %xmm0
movaps %xmm0, 0x600(%rsp)
movl $0x1d, 0x5fc(%rsp)
movaps 0x600(%rsp), %xmm0
movd 0x5fc(%rsp), %xmm1
pslld %xmm1, %xmm0
movaps %xmm0, -0x20(%rsp)
movaps -0x30(%rsp), %xmm0
movaps %xmm0, 0x640(%rsp)
movaps 0x7da211(%rip), %xmm0 # 0x1e1ca30
movaps %xmm0, 0x630(%rsp)
movaps 0x640(%rsp), %xmm0
movaps 0x630(%rsp), %xmm1
pand %xmm1, %xmm0
movaps %xmm0, -0x30(%rsp)
movaps -0x30(%rsp), %xmm1
xorps %xmm0, %xmm0
movaps %xmm0, 0x6b0(%rsp)
movaps 0x6b0(%rsp), %xmm0
movaps %xmm1, 0x6a0(%rsp)
movaps %xmm0, 0x690(%rsp)
movaps 0x6a0(%rsp), %xmm0
movaps 0x690(%rsp), %xmm1
pcmpeqd %xmm1, %xmm0
movaps %xmm0, -0x30(%rsp)
movaps -0x20(%rsp), %xmm0
movaps %xmm0, 0x620(%rsp)
movaps 0x620(%rsp), %xmm0
movaps %xmm0, -0x40(%rsp)
movaps -0x30(%rsp), %xmm0
movaps %xmm0, 0x610(%rsp)
movaps 0x610(%rsp), %xmm0
movaps %xmm0, -0x50(%rsp)
movaps (%rsp), %xmm1
movaps -0x40(%rsp), %xmm0
movaps %xmm1, 0x6f0(%rsp)
movaps %xmm0, 0x6e0(%rsp)
movaps 0x6f0(%rsp), %xmm0
movaps 0x6e0(%rsp), %xmm1
pxor %xmm1, %xmm0
movaps %xmm0, (%rsp)
movaps 0x7da153(%rip), %xmm0 # 0x1e1ca40
movaps %xmm0, 0x30(%rsp)
movaps 0x7da157(%rip), %xmm0 # 0x1e1ca50
movaps %xmm0, 0x20(%rsp)
movaps 0x7da15b(%rip), %xmm0 # 0x1e1ca60
movaps %xmm0, 0x10(%rsp)
movaps -0x10(%rsp), %xmm2
movaps 0x30(%rsp), %xmm1
movaps 0x40(%rsp), %xmm0
movaps %xmm2, 0x580(%rsp)
movaps %xmm1, 0x570(%rsp)
movaps %xmm0, 0x560(%rsp)
movaps 0x580(%rsp), %xmm1
movaps 0x570(%rsp), %xmm0
movaps %xmm1, 0x5a0(%rsp)
movaps %xmm0, 0x590(%rsp)
movaps 0x5a0(%rsp), %xmm1
movaps 0x590(%rsp), %xmm0
mulps %xmm0, %xmm1
movaps 0x560(%rsp), %xmm0
movaps %xmm1, 0x5c0(%rsp)
movaps %xmm0, 0x5b0(%rsp)
movaps 0x5c0(%rsp), %xmm0
movaps 0x5b0(%rsp), %xmm1
addps %xmm1, %xmm0
movaps %xmm0, 0x40(%rsp)
movaps -0x10(%rsp), %xmm2
movaps 0x20(%rsp), %xmm1
movaps 0x40(%rsp), %xmm0
movaps %xmm2, 0x510(%rsp)
movaps %xmm1, 0x500(%rsp)
movaps %xmm0, 0x4f0(%rsp)
movaps 0x510(%rsp), %xmm1
movaps 0x500(%rsp), %xmm0
movaps %xmm1, 0x530(%rsp)
movaps %xmm0, 0x520(%rsp)
movaps 0x530(%rsp), %xmm1
movaps 0x520(%rsp), %xmm0
mulps %xmm0, %xmm1
movaps 0x4f0(%rsp), %xmm0
movaps %xmm1, 0x550(%rsp)
movaps %xmm0, 0x540(%rsp)
movaps 0x550(%rsp), %xmm0
movaps 0x540(%rsp), %xmm1
addps %xmm1, %xmm0
movaps %xmm0, 0x40(%rsp)
movaps -0x10(%rsp), %xmm2
movaps 0x10(%rsp), %xmm1
movaps 0x40(%rsp), %xmm0
movaps %xmm2, 0x4a0(%rsp)
movaps %xmm1, 0x490(%rsp)
movaps %xmm0, 0x480(%rsp)
movaps 0x4a0(%rsp), %xmm1
movaps 0x490(%rsp), %xmm0
movaps %xmm1, 0x4c0(%rsp)
movaps %xmm0, 0x4b0(%rsp)
movaps 0x4c0(%rsp), %xmm1
movaps 0x4b0(%rsp), %xmm0
mulps %xmm0, %xmm1
movaps 0x480(%rsp), %xmm0
movaps %xmm1, 0x4e0(%rsp)
movaps %xmm0, 0x4d0(%rsp)
movaps 0x4e0(%rsp), %xmm0
movaps 0x4d0(%rsp), %xmm1
addps %xmm1, %xmm0
movaps %xmm0, 0x40(%rsp)
movaps 0x7d9fc1(%rip), %xmm0 # 0x1e1ca70
movaps %xmm0, -0x10(%rsp)
movaps 0x40(%rsp), %xmm0
movaps %xmm0, 0x170(%rsp)
movaps %xmm0, 0x160(%rsp)
movaps 0x170(%rsp), %xmm0
movaps 0x160(%rsp), %xmm1
mulps %xmm1, %xmm0
movaps %xmm0, -0x60(%rsp)
movaps -0x10(%rsp), %xmm1
movaps -0x60(%rsp), %xmm0
movaps %xmm1, 0x430(%rsp)
movaps %xmm0, 0x420(%rsp)
movaps 0x7d9f7e(%rip), %xmm0 # 0x1e1ca80
movaps %xmm0, 0x410(%rsp)
movaps 0x430(%rsp), %xmm1
movaps 0x420(%rsp), %xmm0
movaps %xmm1, 0x450(%rsp)
movaps %xmm0, 0x440(%rsp)
movaps 0x450(%rsp), %xmm1
movaps 0x440(%rsp), %xmm0
mulps %xmm0, %xmm1
movaps 0x410(%rsp), %xmm0
movaps %xmm1, 0x470(%rsp)
movaps %xmm0, 0x460(%rsp)
movaps 0x470(%rsp), %xmm0
movaps 0x460(%rsp), %xmm1
addps %xmm1, %xmm0
movaps %xmm0, -0x10(%rsp)
movaps -0x10(%rsp), %xmm1
movaps -0x60(%rsp), %xmm0
movaps %xmm1, 0x3c0(%rsp)
movaps %xmm0, 0x3b0(%rsp)
movaps 0x7d9f02(%rip), %xmm0 # 0x1e1ca90
movaps %xmm0, 0x3a0(%rsp)
movaps 0x3c0(%rsp), %xmm1
movaps 0x3b0(%rsp), %xmm0
movaps %xmm1, 0x3e0(%rsp)
movaps %xmm0, 0x3d0(%rsp)
movaps 0x3e0(%rsp), %xmm1
movaps 0x3d0(%rsp), %xmm0
mulps %xmm0, %xmm1
movaps 0x3a0(%rsp), %xmm0
movaps %xmm1, 0x400(%rsp)
movaps %xmm0, 0x3f0(%rsp)
movaps 0x400(%rsp), %xmm0
movaps 0x3f0(%rsp), %xmm1
addps %xmm1, %xmm0
movaps %xmm0, -0x10(%rsp)
movaps -0x10(%rsp), %xmm1
movaps -0x60(%rsp), %xmm0
movaps %xmm1, 0x150(%rsp)
movaps %xmm0, 0x140(%rsp)
movaps 0x150(%rsp), %xmm0
movaps 0x140(%rsp), %xmm1
mulps %xmm1, %xmm0
movaps %xmm0, -0x10(%rsp)
movaps -0x10(%rsp), %xmm1
movaps -0x60(%rsp), %xmm0
movaps %xmm1, 0x130(%rsp)
movaps %xmm0, 0x120(%rsp)
movaps 0x130(%rsp), %xmm0
movaps 0x120(%rsp), %xmm1
mulps %xmm1, %xmm0
movaps %xmm0, -0x10(%rsp)
movaps -0x60(%rsp), %xmm1
movaps -0x10(%rsp), %xmm0
movaps %xmm1, 0x200(%rsp)
movaps 0x7c8e0a(%rip), %xmm1 # 0x1e0ba80
movaps %xmm1, 0x1f0(%rsp)
movaps %xmm0, 0x1e0(%rsp)
movaps 0x1e0(%rsp), %xmm1
movaps 0x200(%rsp), %xmm2
movaps 0x1f0(%rsp), %xmm0
movaps %xmm2, 0x240(%rsp)
movaps %xmm0, 0x230(%rsp)
movaps 0x240(%rsp), %xmm0
movaps 0x230(%rsp), %xmm2
mulps %xmm2, %xmm0
movaps %xmm1, 0x220(%rsp)
movaps %xmm0, 0x210(%rsp)
movaps 0x220(%rsp), %xmm0
movaps 0x210(%rsp), %xmm1
subps %xmm1, %xmm0
movaps %xmm0, -0x10(%rsp)
movaps -0x10(%rsp), %xmm0
movaps %xmm0, 0x1d0(%rsp)
movaps 0x7c8d43(%rip), %xmm0 # 0x1e0ba40
movaps %xmm0, 0x1c0(%rsp)
movaps 0x1d0(%rsp), %xmm0
movaps 0x1c0(%rsp), %xmm1
addps %xmm1, %xmm0
movaps %xmm0, -0x10(%rsp)
movaps 0x7d9d7c(%rip), %xmm0 # 0x1e1caa0
movaps %xmm0, -0x70(%rsp)
movaps -0x70(%rsp), %xmm1
movaps -0x60(%rsp), %xmm0
movaps %xmm1, 0x350(%rsp)
movaps %xmm0, 0x340(%rsp)
movaps 0x7d9d66(%rip), %xmm0 # 0x1e1cab0
movaps %xmm0, 0x330(%rsp)
movaps 0x350(%rsp), %xmm1
movaps 0x340(%rsp), %xmm0
movaps %xmm1, 0x370(%rsp)
movaps %xmm0, 0x360(%rsp)
movaps 0x370(%rsp), %xmm1
movaps 0x360(%rsp), %xmm0
mulps %xmm0, %xmm1
movaps 0x330(%rsp), %xmm0
movaps %xmm1, 0x390(%rsp)
movaps %xmm0, 0x380(%rsp)
movaps 0x390(%rsp), %xmm0
movaps 0x380(%rsp), %xmm1
addps %xmm1, %xmm0
movaps %xmm0, -0x70(%rsp)
movaps -0x70(%rsp), %xmm1
movaps -0x60(%rsp), %xmm0
movaps %xmm1, 0x2e0(%rsp)
movaps %xmm0, 0x2d0(%rsp)
movaps 0x7d9cea(%rip), %xmm0 # 0x1e1cac0
movaps %xmm0, 0x2c0(%rsp)
movaps 0x2e0(%rsp), %xmm1
movaps 0x2d0(%rsp), %xmm0
movaps %xmm1, 0x300(%rsp)
movaps %xmm0, 0x2f0(%rsp)
movaps 0x300(%rsp), %xmm1
movaps 0x2f0(%rsp), %xmm0
mulps %xmm0, %xmm1
movaps 0x2c0(%rsp), %xmm0
movaps %xmm1, 0x320(%rsp)
movaps %xmm0, 0x310(%rsp)
movaps 0x320(%rsp), %xmm0
movaps 0x310(%rsp), %xmm1
addps %xmm1, %xmm0
movaps %xmm0, -0x70(%rsp)
movaps -0x70(%rsp), %xmm1
movaps -0x60(%rsp), %xmm0
movaps %xmm1, 0x110(%rsp)
movaps %xmm0, 0x100(%rsp)
movaps 0x110(%rsp), %xmm0
movaps 0x100(%rsp), %xmm1
mulps %xmm1, %xmm0
movaps %xmm0, -0x70(%rsp)
movaps -0x70(%rsp), %xmm1
movaps 0x40(%rsp), %xmm0
movaps %xmm1, 0x270(%rsp)
movaps %xmm0, 0x260(%rsp)
movaps %xmm0, 0x250(%rsp)
movaps 0x270(%rsp), %xmm1
movaps 0x260(%rsp), %xmm0
movaps %xmm1, 0x290(%rsp)
movaps %xmm0, 0x280(%rsp)
movaps 0x290(%rsp), %xmm1
movaps 0x280(%rsp), %xmm0
mulps %xmm0, %xmm1
movaps 0x250(%rsp), %xmm0
movaps %xmm1, 0x2b0(%rsp)
movaps %xmm0, 0x2a0(%rsp)
movaps 0x2b0(%rsp), %xmm0
movaps 0x2a0(%rsp), %xmm1
addps %xmm1, %xmm0
movaps %xmm0, -0x70(%rsp)
movaps -0x50(%rsp), %xmm0
movaps %xmm0, 0x10(%rsp)
movaps 0x10(%rsp), %xmm1
movaps -0x70(%rsp), %xmm0
movaps %xmm1, 0x90(%rsp)
movaps %xmm0, 0x80(%rsp)
movaps 0x90(%rsp), %xmm0
movaps 0x80(%rsp), %xmm1
pand %xmm1, %xmm0
movaps %xmm0, -0x70(%rsp)
movaps 0x10(%rsp), %xmm1
movaps -0x10(%rsp), %xmm0
movaps %xmm1, 0x60(%rsp)
movaps %xmm0, 0x50(%rsp)
movaps 0x60(%rsp), %xmm0
pcmpeqd %xmm1, %xmm1
pxor %xmm1, %xmm0
movaps 0x50(%rsp), %xmm1
pand %xmm1, %xmm0
movaps %xmm0, -0x10(%rsp)
movaps -0x10(%rsp), %xmm1
movaps -0x70(%rsp), %xmm0
movaps %xmm1, 0x1b0(%rsp)
movaps %xmm0, 0x1a0(%rsp)
movaps 0x1b0(%rsp), %xmm0
addps 0x1a0(%rsp), %xmm0
movaps %xmm0, -0x10(%rsp)
movaps -0x10(%rsp), %xmm1
movaps (%rsp), %xmm0
movaps %xmm1, 0x6d0(%rsp)
movaps %xmm0, 0x6c0(%rsp)
movaps 0x6d0(%rsp), %xmm0
movaps 0x6c0(%rsp), %xmm1
pxor %xmm1, %xmm0
movaps %xmm0, -0x10(%rsp)
movaps -0x10(%rsp), %xmm0
addq $0x708, %rsp # imm = 0x708
retq
nopw %cs:(%rax,%rax)
nopl (%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,217 |
ncnn::UnaryOp_x86_functor::unary_op_sin::func(float const&) const
|
float func(const float& x) const
{
return (float)sin(x);
}
|
subq $0x18, %rsp
movq %rdi, 0x10(%rsp)
movq %rsi, 0x8(%rsp)
movq 0x8(%rsp), %rax
movss (%rax), %xmm0
callq 0x163a260
addq $0x18, %rsp
retq
nopw %cs:(%rax,%rax)
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,218 |
ncnn::UnaryOp_x86_functor::unary_op_cos::func_pack4(float vector[4] const&) const
|
__m128 func_pack4(const __m128& x) const
{
return cos_ps(x);
}
|
subq $0x6d8, %rsp # imm = 0x6D8
movq %rdi, -0x78(%rsp)
movq %rsi, -0x80(%rsp)
movq -0x80(%rsp), %rax
movaps (%rax), %xmm0
movaps %xmm0, 0x30(%rsp)
xorps %xmm0, %xmm0
movaps %xmm0, 0x60(%rsp)
movaps 0x60(%rsp), %xmm0
movaps %xmm0, 0x10(%rsp)
movaps 0x30(%rsp), %xmm0
movaps %xmm0, 0xa0(%rsp)
movaps 0x7d50cc(%rip), %xmm0 # 0x1e18120
movaps %xmm0, 0x90(%rsp)
movaps 0xa0(%rsp), %xmm0
movaps 0x90(%rsp), %xmm1
pand %xmm1, %xmm0
movaps %xmm0, 0x30(%rsp)
movaps 0x30(%rsp), %xmm0
movaps %xmm0, 0x160(%rsp)
movaps 0x7d9967(%rip), %xmm0 # 0x1e1c9f0
movaps %xmm0, 0x150(%rsp)
movaps 0x160(%rsp), %xmm0
movaps 0x150(%rsp), %xmm1
mulps %xmm1, %xmm0
movaps %xmm0, -0x10(%rsp)
movaps -0x10(%rsp), %xmm0
movaps %xmm0, 0xc0(%rsp)
cvttps2dq 0xc0(%rsp), %xmm0
movaps %xmm0, -0x30(%rsp)
movaps -0x30(%rsp), %xmm0
movaps %xmm0, 0x5b0(%rsp)
movaps 0x7d9928(%rip), %xmm0 # 0x1e1ca00
movaps %xmm0, 0x5a0(%rsp)
movaps 0x5b0(%rsp), %xmm0
movaps 0x5a0(%rsp), %xmm1
paddd %xmm1, %xmm0
movaps %xmm0, -0x30(%rsp)
movaps -0x30(%rsp), %xmm0
movaps %xmm0, 0x650(%rsp)
movaps 0x7d9903(%rip), %xmm0 # 0x1e1ca10
movaps %xmm0, 0x640(%rsp)
movaps 0x650(%rsp), %xmm0
movaps 0x640(%rsp), %xmm1
pand %xmm1, %xmm0
movaps %xmm0, -0x30(%rsp)
movaps -0x30(%rsp), %xmm0
movaps %xmm0, 0xb0(%rsp)
cvtdq2ps 0xb0(%rsp), %xmm0
movaps %xmm0, -0x10(%rsp)
movaps -0x30(%rsp), %xmm0
movaps %xmm0, 0x610(%rsp)
movaps 0x7d98d4(%rip), %xmm0 # 0x1e1ca30
movaps %xmm0, 0x600(%rsp)
movaps 0x610(%rsp), %xmm1
movaps 0x600(%rsp), %xmm2
psubd %xmm2, %xmm1
movaps %xmm1, -0x30(%rsp)
movaps -0x30(%rsp), %xmm1
movaps %xmm1, 0x6c0(%rsp)
movaps 0x7d988f(%rip), %xmm1 # 0x1e1ca20
movaps %xmm1, 0x6b0(%rsp)
movaps 0x6c0(%rsp), %xmm1
movaps 0x6b0(%rsp), %xmm2
pandn %xmm2, %xmm1
movaps %xmm1, -0x20(%rsp)
movaps -0x20(%rsp), %xmm1
movaps %xmm1, 0x5d0(%rsp)
movl $0x1d, 0x5cc(%rsp)
movaps 0x5d0(%rsp), %xmm1
movd 0x5cc(%rsp), %xmm2
pslld %xmm2, %xmm1
movaps %xmm1, -0x20(%rsp)
movaps -0x30(%rsp), %xmm1
movaps %xmm1, 0x630(%rsp)
movaps %xmm0, 0x620(%rsp)
movaps 0x630(%rsp), %xmm0
movaps 0x620(%rsp), %xmm1
pand %xmm1, %xmm0
movaps %xmm0, -0x30(%rsp)
movaps -0x30(%rsp), %xmm1
xorps %xmm0, %xmm0
movaps %xmm0, 0x680(%rsp)
movaps 0x680(%rsp), %xmm0
movaps %xmm1, 0x670(%rsp)
movaps %xmm0, 0x660(%rsp)
movaps 0x670(%rsp), %xmm0
movaps 0x660(%rsp), %xmm1
pcmpeqd %xmm1, %xmm0
movaps %xmm0, -0x30(%rsp)
movaps -0x20(%rsp), %xmm0
movaps %xmm0, 0x5f0(%rsp)
movaps 0x5f0(%rsp), %xmm0
movaps %xmm0, -0x40(%rsp)
movaps -0x30(%rsp), %xmm0
movaps %xmm0, 0x5e0(%rsp)
movaps 0x5e0(%rsp), %xmm0
movaps %xmm0, -0x50(%rsp)
movaps 0x7d97b2(%rip), %xmm0 # 0x1e1ca40
movaps %xmm0, 0x20(%rsp)
movaps 0x7d97b6(%rip), %xmm0 # 0x1e1ca50
movaps %xmm0, 0x10(%rsp)
movaps 0x7d97ba(%rip), %xmm0 # 0x1e1ca60
movaps %xmm0, (%rsp)
movaps -0x10(%rsp), %xmm2
movaps 0x20(%rsp), %xmm1
movaps 0x30(%rsp), %xmm0
movaps %xmm2, 0x550(%rsp)
movaps %xmm1, 0x540(%rsp)
movaps %xmm0, 0x530(%rsp)
movaps 0x550(%rsp), %xmm1
movaps 0x540(%rsp), %xmm0
movaps %xmm1, 0x570(%rsp)
movaps %xmm0, 0x560(%rsp)
movaps 0x570(%rsp), %xmm1
movaps 0x560(%rsp), %xmm0
mulps %xmm0, %xmm1
movaps 0x530(%rsp), %xmm0
movaps %xmm1, 0x590(%rsp)
movaps %xmm0, 0x580(%rsp)
movaps 0x590(%rsp), %xmm0
movaps 0x580(%rsp), %xmm1
addps %xmm1, %xmm0
movaps %xmm0, 0x30(%rsp)
movaps -0x10(%rsp), %xmm2
movaps 0x10(%rsp), %xmm1
movaps 0x30(%rsp), %xmm0
movaps %xmm2, 0x4e0(%rsp)
movaps %xmm1, 0x4d0(%rsp)
movaps %xmm0, 0x4c0(%rsp)
movaps 0x4e0(%rsp), %xmm1
movaps 0x4d0(%rsp), %xmm0
movaps %xmm1, 0x500(%rsp)
movaps %xmm0, 0x4f0(%rsp)
movaps 0x500(%rsp), %xmm1
movaps 0x4f0(%rsp), %xmm0
mulps %xmm0, %xmm1
movaps 0x4c0(%rsp), %xmm0
movaps %xmm1, 0x520(%rsp)
movaps %xmm0, 0x510(%rsp)
movaps 0x520(%rsp), %xmm0
movaps 0x510(%rsp), %xmm1
addps %xmm1, %xmm0
movaps %xmm0, 0x30(%rsp)
movaps -0x10(%rsp), %xmm2
movaps (%rsp), %xmm1
movaps 0x30(%rsp), %xmm0
movaps %xmm2, 0x470(%rsp)
movaps %xmm1, 0x460(%rsp)
movaps %xmm0, 0x450(%rsp)
movaps 0x470(%rsp), %xmm1
movaps 0x460(%rsp), %xmm0
movaps %xmm1, 0x490(%rsp)
movaps %xmm0, 0x480(%rsp)
movaps 0x490(%rsp), %xmm1
movaps 0x480(%rsp), %xmm0
mulps %xmm0, %xmm1
movaps 0x450(%rsp), %xmm0
movaps %xmm1, 0x4b0(%rsp)
movaps %xmm0, 0x4a0(%rsp)
movaps 0x4b0(%rsp), %xmm0
movaps 0x4a0(%rsp), %xmm1
addps %xmm1, %xmm0
movaps %xmm0, 0x30(%rsp)
movaps 0x7d9622(%rip), %xmm0 # 0x1e1ca70
movaps %xmm0, -0x10(%rsp)
movaps 0x30(%rsp), %xmm0
movaps %xmm0, 0x140(%rsp)
movaps %xmm0, 0x130(%rsp)
movaps 0x140(%rsp), %xmm0
movaps 0x130(%rsp), %xmm1
mulps %xmm1, %xmm0
movaps %xmm0, -0x60(%rsp)
movaps -0x10(%rsp), %xmm1
movaps -0x60(%rsp), %xmm0
movaps %xmm1, 0x400(%rsp)
movaps %xmm0, 0x3f0(%rsp)
movaps 0x7d95df(%rip), %xmm0 # 0x1e1ca80
movaps %xmm0, 0x3e0(%rsp)
movaps 0x400(%rsp), %xmm1
movaps 0x3f0(%rsp), %xmm0
movaps %xmm1, 0x420(%rsp)
movaps %xmm0, 0x410(%rsp)
movaps 0x420(%rsp), %xmm1
movaps 0x410(%rsp), %xmm0
mulps %xmm0, %xmm1
movaps 0x3e0(%rsp), %xmm0
movaps %xmm1, 0x440(%rsp)
movaps %xmm0, 0x430(%rsp)
movaps 0x440(%rsp), %xmm0
movaps 0x430(%rsp), %xmm1
addps %xmm1, %xmm0
movaps %xmm0, -0x10(%rsp)
movaps -0x10(%rsp), %xmm1
movaps -0x60(%rsp), %xmm0
movaps %xmm1, 0x390(%rsp)
movaps %xmm0, 0x380(%rsp)
movaps 0x7d9563(%rip), %xmm0 # 0x1e1ca90
movaps %xmm0, 0x370(%rsp)
movaps 0x390(%rsp), %xmm1
movaps 0x380(%rsp), %xmm0
movaps %xmm1, 0x3b0(%rsp)
movaps %xmm0, 0x3a0(%rsp)
movaps 0x3b0(%rsp), %xmm1
movaps 0x3a0(%rsp), %xmm0
mulps %xmm0, %xmm1
movaps 0x370(%rsp), %xmm0
movaps %xmm1, 0x3d0(%rsp)
movaps %xmm0, 0x3c0(%rsp)
movaps 0x3d0(%rsp), %xmm0
movaps 0x3c0(%rsp), %xmm1
addps %xmm1, %xmm0
movaps %xmm0, -0x10(%rsp)
movaps -0x10(%rsp), %xmm1
movaps -0x60(%rsp), %xmm0
movaps %xmm1, 0x120(%rsp)
movaps %xmm0, 0x110(%rsp)
movaps 0x120(%rsp), %xmm0
movaps 0x110(%rsp), %xmm1
mulps %xmm1, %xmm0
movaps %xmm0, -0x10(%rsp)
movaps -0x10(%rsp), %xmm1
movaps -0x60(%rsp), %xmm0
movaps %xmm1, 0x100(%rsp)
movaps %xmm0, 0xf0(%rsp)
movaps 0x100(%rsp), %xmm0
movaps 0xf0(%rsp), %xmm1
mulps %xmm1, %xmm0
movaps %xmm0, -0x10(%rsp)
movaps -0x60(%rsp), %xmm1
movaps -0x10(%rsp), %xmm0
movaps %xmm1, 0x1d0(%rsp)
movaps 0x7c846b(%rip), %xmm1 # 0x1e0ba80
movaps %xmm1, 0x1c0(%rsp)
movaps %xmm0, 0x1b0(%rsp)
movaps 0x1b0(%rsp), %xmm1
movaps 0x1d0(%rsp), %xmm2
movaps 0x1c0(%rsp), %xmm0
movaps %xmm2, 0x210(%rsp)
movaps %xmm0, 0x200(%rsp)
movaps 0x210(%rsp), %xmm0
movaps 0x200(%rsp), %xmm2
mulps %xmm2, %xmm0
movaps %xmm1, 0x1f0(%rsp)
movaps %xmm0, 0x1e0(%rsp)
movaps 0x1f0(%rsp), %xmm0
movaps 0x1e0(%rsp), %xmm1
subps %xmm1, %xmm0
movaps %xmm0, -0x10(%rsp)
movaps -0x10(%rsp), %xmm0
movaps %xmm0, 0x1a0(%rsp)
movaps 0x7c83a4(%rip), %xmm0 # 0x1e0ba40
movaps %xmm0, 0x190(%rsp)
movaps 0x1a0(%rsp), %xmm0
movaps 0x190(%rsp), %xmm1
addps %xmm1, %xmm0
movaps %xmm0, -0x10(%rsp)
movaps 0x7d93dd(%rip), %xmm0 # 0x1e1caa0
movaps %xmm0, -0x70(%rsp)
movaps -0x70(%rsp), %xmm1
movaps -0x60(%rsp), %xmm0
movaps %xmm1, 0x320(%rsp)
movaps %xmm0, 0x310(%rsp)
movaps 0x7d93c7(%rip), %xmm0 # 0x1e1cab0
movaps %xmm0, 0x300(%rsp)
movaps 0x320(%rsp), %xmm1
movaps 0x310(%rsp), %xmm0
movaps %xmm1, 0x340(%rsp)
movaps %xmm0, 0x330(%rsp)
movaps 0x340(%rsp), %xmm1
movaps 0x330(%rsp), %xmm0
mulps %xmm0, %xmm1
movaps 0x300(%rsp), %xmm0
movaps %xmm1, 0x360(%rsp)
movaps %xmm0, 0x350(%rsp)
movaps 0x360(%rsp), %xmm0
movaps 0x350(%rsp), %xmm1
addps %xmm1, %xmm0
movaps %xmm0, -0x70(%rsp)
movaps -0x70(%rsp), %xmm1
movaps -0x60(%rsp), %xmm0
movaps %xmm1, 0x2b0(%rsp)
movaps %xmm0, 0x2a0(%rsp)
movaps 0x7d934b(%rip), %xmm0 # 0x1e1cac0
movaps %xmm0, 0x290(%rsp)
movaps 0x2b0(%rsp), %xmm1
movaps 0x2a0(%rsp), %xmm0
movaps %xmm1, 0x2d0(%rsp)
movaps %xmm0, 0x2c0(%rsp)
movaps 0x2d0(%rsp), %xmm1
movaps 0x2c0(%rsp), %xmm0
mulps %xmm0, %xmm1
movaps 0x290(%rsp), %xmm0
movaps %xmm1, 0x2f0(%rsp)
movaps %xmm0, 0x2e0(%rsp)
movaps 0x2f0(%rsp), %xmm0
movaps 0x2e0(%rsp), %xmm1
addps %xmm1, %xmm0
movaps %xmm0, -0x70(%rsp)
movaps -0x70(%rsp), %xmm1
movaps -0x60(%rsp), %xmm0
movaps %xmm1, 0xe0(%rsp)
movaps %xmm0, 0xd0(%rsp)
movaps 0xe0(%rsp), %xmm0
movaps 0xd0(%rsp), %xmm1
mulps %xmm1, %xmm0
movaps %xmm0, -0x70(%rsp)
movaps -0x70(%rsp), %xmm1
movaps 0x30(%rsp), %xmm0
movaps %xmm1, 0x240(%rsp)
movaps %xmm0, 0x230(%rsp)
movaps %xmm0, 0x220(%rsp)
movaps 0x240(%rsp), %xmm1
movaps 0x230(%rsp), %xmm0
movaps %xmm1, 0x260(%rsp)
movaps %xmm0, 0x250(%rsp)
movaps 0x260(%rsp), %xmm1
movaps 0x250(%rsp), %xmm0
mulps %xmm0, %xmm1
movaps 0x220(%rsp), %xmm0
movaps %xmm1, 0x280(%rsp)
movaps %xmm0, 0x270(%rsp)
movaps 0x280(%rsp), %xmm0
movaps 0x270(%rsp), %xmm1
addps %xmm1, %xmm0
movaps %xmm0, -0x70(%rsp)
movaps -0x50(%rsp), %xmm0
movaps %xmm0, (%rsp)
movaps (%rsp), %xmm1
movaps -0x70(%rsp), %xmm0
movaps %xmm1, 0x80(%rsp)
movaps %xmm0, 0x70(%rsp)
movaps 0x80(%rsp), %xmm0
movaps 0x70(%rsp), %xmm1
pand %xmm1, %xmm0
movaps %xmm0, -0x70(%rsp)
movaps (%rsp), %xmm1
movaps -0x10(%rsp), %xmm0
movaps %xmm1, 0x50(%rsp)
movaps %xmm0, 0x40(%rsp)
movaps 0x50(%rsp), %xmm0
pcmpeqd %xmm1, %xmm1
pxor %xmm1, %xmm0
movaps 0x40(%rsp), %xmm1
pand %xmm1, %xmm0
movaps %xmm0, -0x10(%rsp)
movaps -0x10(%rsp), %xmm1
movaps -0x70(%rsp), %xmm0
movaps %xmm1, 0x180(%rsp)
movaps %xmm0, 0x170(%rsp)
movaps 0x180(%rsp), %xmm0
addps 0x170(%rsp), %xmm0
movaps %xmm0, -0x10(%rsp)
movaps -0x10(%rsp), %xmm1
movaps -0x40(%rsp), %xmm0
movaps %xmm1, 0x6a0(%rsp)
movaps %xmm0, 0x690(%rsp)
movaps 0x6a0(%rsp), %xmm0
movaps 0x690(%rsp), %xmm1
pxor %xmm1, %xmm0
movaps %xmm0, -0x10(%rsp)
movaps -0x10(%rsp), %xmm0
addq $0x6d8, %rsp # imm = 0x6D8
retq
nopl (%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,219 |
ncnn::UnaryOp_x86_functor::unary_op_cos::func(float const&) const
|
float func(const float& x) const
{
return (float)cos(x);
}
|
subq $0x18, %rsp
movq %rdi, 0x10(%rsp)
movq %rsi, 0x8(%rsp)
movq 0x8(%rsp), %rax
movss (%rax), %xmm0
callq 0x163a2b0
addq $0x18, %rsp
retq
nopw %cs:(%rax,%rax)
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,220 |
ncnn::UnaryOp_x86_functor::unary_op_tan::func_pack4(float vector[4] const&) const
|
__m128 func_pack4(const __m128& x) const
{
return tan_ps(x);
}
|
subq $0x968, %rsp # imm = 0x968
movq %rdi, -0x78(%rsp)
movq %rsi, -0x80(%rsp)
movq -0x80(%rsp), %rax
movaps (%rax), %xmm0
movaps %xmm0, -0x10(%rsp)
movl $0x322bcc77, 0x1c(%rsp) # imm = 0x322BCC77
movss 0x1c(%rsp), %xmm0
shufps $0x0, %xmm0, %xmm0 # xmm0 = xmm0[0,0,0,0]
movaps %xmm0, (%rsp)
movaps (%rsp), %xmm0
movaps %xmm0, -0x40(%rsp)
movaps -0x10(%rsp), %xmm0
movaps %xmm0, 0x170(%rsp)
leaq -0x20(%rsp), %rax
movq %rax, 0x168(%rsp)
leaq -0x30(%rsp), %rax
movq %rax, 0x160(%rsp)
xorps %xmm0, %xmm0
movaps %xmm0, 0x1e0(%rsp)
movaps 0x1e0(%rsp), %xmm1
movaps %xmm1, 0x130(%rsp)
movaps 0x170(%rsp), %xmm1
movaps %xmm1, 0x120(%rsp)
movaps 0x170(%rsp), %xmm1
movaps %xmm1, 0x240(%rsp)
movaps 0x7d46da(%rip), %xmm1 # 0x1e18120
movaps %xmm1, 0x230(%rsp)
movaps 0x240(%rsp), %xmm1
movaps 0x230(%rsp), %xmm2
pand %xmm2, %xmm1
movaps %xmm1, 0x170(%rsp)
movaps 0x120(%rsp), %xmm1
movaps %xmm1, 0x220(%rsp)
movaps 0x7cc0df(%rip), %xmm1 # 0x1e0fb60
movaps %xmm1, 0x210(%rsp)
movaps 0x220(%rsp), %xmm1
movaps 0x210(%rsp), %xmm2
pand %xmm2, %xmm1
movaps %xmm1, 0x120(%rsp)
movaps 0x170(%rsp), %xmm1
movaps %xmm1, 0x300(%rsp)
movaps 0x7d8f34(%rip), %xmm1 # 0x1e1c9f0
movaps %xmm1, 0x2f0(%rsp)
movaps 0x300(%rsp), %xmm1
movaps 0x2f0(%rsp), %xmm2
mulps %xmm2, %xmm1
movaps %xmm1, 0x110(%rsp)
movaps 0x110(%rsp), %xmm1
movaps %xmm1, 0x260(%rsp)
cvttps2dq 0x260(%rsp), %xmm1
movaps %xmm1, 0xf0(%rsp)
movaps 0xf0(%rsp), %xmm1
movaps %xmm1, 0x770(%rsp)
movaps 0x7d8ee9(%rip), %xmm1 # 0x1e1ca00
movaps %xmm1, 0x760(%rsp)
movaps 0x770(%rsp), %xmm1
movaps 0x760(%rsp), %xmm2
paddd %xmm2, %xmm1
movaps %xmm1, 0xf0(%rsp)
movaps 0xf0(%rsp), %xmm1
movaps %xmm1, 0x860(%rsp)
movaps 0x7d8ebe(%rip), %xmm1 # 0x1e1ca10
movaps %xmm1, 0x850(%rsp)
movaps 0x860(%rsp), %xmm1
movaps 0x850(%rsp), %xmm2
pand %xmm2, %xmm1
movaps %xmm1, 0xf0(%rsp)
movaps 0xf0(%rsp), %xmm1
movaps %xmm1, 0x250(%rsp)
cvtdq2ps 0x250(%rsp), %xmm1
movaps %xmm1, 0x110(%rsp)
movaps 0xf0(%rsp), %xmm1
movaps %xmm1, 0xe0(%rsp)
movaps 0xf0(%rsp), %xmm1
movaps %xmm1, 0x840(%rsp)
movaps 0x7d8e63(%rip), %xmm1 # 0x1e1ca20
movaps %xmm1, 0x830(%rsp)
movaps 0x840(%rsp), %xmm2
movaps 0x830(%rsp), %xmm3
pand %xmm3, %xmm2
movaps %xmm2, 0x100(%rsp)
movaps 0x100(%rsp), %xmm2
movaps %xmm2, 0x7b0(%rsp)
movl $0x1d, 0x7ac(%rsp)
movaps 0x7b0(%rsp), %xmm2
movd 0x7ac(%rsp), %xmm3
pslld %xmm3, %xmm2
movaps %xmm2, 0x100(%rsp)
movaps 0x100(%rsp), %xmm2
movaps %xmm2, 0x7e0(%rsp)
movaps 0x7e0(%rsp), %xmm2
movaps %xmm2, 0xd0(%rsp)
movaps 0xf0(%rsp), %xmm2
movaps %xmm2, 0x820(%rsp)
movaps 0x7d8de0(%rip), %xmm2 # 0x1e1ca30
movaps %xmm2, 0x810(%rsp)
movaps 0x820(%rsp), %xmm3
movaps 0x810(%rsp), %xmm4
pand %xmm4, %xmm3
movaps %xmm3, 0xf0(%rsp)
movaps 0xf0(%rsp), %xmm4
xorps %xmm3, %xmm3
movaps %xmm3, 0x890(%rsp)
movaps 0x890(%rsp), %xmm3
movaps %xmm4, 0x880(%rsp)
movaps %xmm3, 0x870(%rsp)
movaps 0x880(%rsp), %xmm3
movaps 0x870(%rsp), %xmm4
pcmpeqd %xmm4, %xmm3
movaps %xmm3, 0xf0(%rsp)
movaps 0xf0(%rsp), %xmm3
movaps %xmm3, 0x7d0(%rsp)
movaps 0x7d0(%rsp), %xmm3
movaps %xmm3, 0xc0(%rsp)
movaps 0x7d8d5e(%rip), %xmm3 # 0x1e1ca40
movaps %xmm3, 0x150(%rsp)
movaps 0x7d8d5f(%rip), %xmm3 # 0x1e1ca50
movaps %xmm3, 0x140(%rsp)
movaps 0x7d8d60(%rip), %xmm3 # 0x1e1ca60
movaps %xmm3, 0x130(%rsp)
movaps 0x110(%rsp), %xmm5
movaps 0x150(%rsp), %xmm4
movaps 0x170(%rsp), %xmm3
movaps %xmm5, 0x710(%rsp)
movaps %xmm4, 0x700(%rsp)
movaps %xmm3, 0x6f0(%rsp)
movaps 0x710(%rsp), %xmm4
movaps 0x700(%rsp), %xmm3
movaps %xmm4, 0x730(%rsp)
movaps %xmm3, 0x720(%rsp)
movaps 0x730(%rsp), %xmm4
movaps 0x720(%rsp), %xmm3
mulps %xmm3, %xmm4
movaps 0x6f0(%rsp), %xmm3
movaps %xmm4, 0x750(%rsp)
movaps %xmm3, 0x740(%rsp)
movaps 0x750(%rsp), %xmm3
movaps 0x740(%rsp), %xmm4
addps %xmm4, %xmm3
movaps %xmm3, 0x170(%rsp)
movaps 0x110(%rsp), %xmm5
movaps 0x140(%rsp), %xmm4
movaps 0x170(%rsp), %xmm3
movaps %xmm5, 0x6a0(%rsp)
movaps %xmm4, 0x690(%rsp)
movaps %xmm3, 0x680(%rsp)
movaps 0x6a0(%rsp), %xmm4
movaps 0x690(%rsp), %xmm3
movaps %xmm4, 0x6c0(%rsp)
movaps %xmm3, 0x6b0(%rsp)
movaps 0x6c0(%rsp), %xmm4
movaps 0x6b0(%rsp), %xmm3
mulps %xmm3, %xmm4
movaps 0x680(%rsp), %xmm3
movaps %xmm4, 0x6e0(%rsp)
movaps %xmm3, 0x6d0(%rsp)
movaps 0x6e0(%rsp), %xmm3
movaps 0x6d0(%rsp), %xmm4
addps %xmm4, %xmm3
movaps %xmm3, 0x170(%rsp)
movaps 0x110(%rsp), %xmm5
movaps 0x130(%rsp), %xmm4
movaps 0x170(%rsp), %xmm3
movaps %xmm5, 0x630(%rsp)
movaps %xmm4, 0x620(%rsp)
movaps %xmm3, 0x610(%rsp)
movaps 0x630(%rsp), %xmm4
movaps 0x620(%rsp), %xmm3
movaps %xmm4, 0x650(%rsp)
movaps %xmm3, 0x640(%rsp)
movaps 0x650(%rsp), %xmm4
movaps 0x640(%rsp), %xmm3
mulps %xmm3, %xmm4
movaps 0x610(%rsp), %xmm3
movaps %xmm4, 0x670(%rsp)
movaps %xmm3, 0x660(%rsp)
movaps 0x670(%rsp), %xmm3
movaps 0x660(%rsp), %xmm4
addps %xmm4, %xmm3
movaps %xmm3, 0x170(%rsp)
movaps 0xe0(%rsp), %xmm3
movaps %xmm3, 0x800(%rsp)
movaps %xmm2, 0x7f0(%rsp)
movaps 0x800(%rsp), %xmm2
movaps 0x7f0(%rsp), %xmm3
psubd %xmm3, %xmm2
movaps %xmm2, 0xe0(%rsp)
movaps 0xe0(%rsp), %xmm2
movaps %xmm2, 0x910(%rsp)
movaps %xmm1, 0x900(%rsp)
movaps 0x910(%rsp), %xmm1
movaps 0x900(%rsp), %xmm2
pandn %xmm2, %xmm1
movaps %xmm1, 0xe0(%rsp)
movaps 0xe0(%rsp), %xmm1
movaps %xmm1, 0x790(%rsp)
movl $0x1d, 0x78c(%rsp)
movaps 0x790(%rsp), %xmm1
movd 0x78c(%rsp), %xmm2
pslld %xmm2, %xmm1
movaps %xmm1, 0xe0(%rsp)
movaps 0xe0(%rsp), %xmm1
movaps %xmm1, 0x7c0(%rsp)
movaps 0x7c0(%rsp), %xmm1
movaps %xmm1, 0xb0(%rsp)
movaps 0x120(%rsp), %xmm2
movaps 0xd0(%rsp), %xmm1
movaps %xmm2, 0x8f0(%rsp)
movaps %xmm1, 0x8e0(%rsp)
movaps 0x8f0(%rsp), %xmm1
movaps 0x8e0(%rsp), %xmm2
pxor %xmm2, %xmm1
movaps %xmm1, 0x120(%rsp)
movaps 0x170(%rsp), %xmm1
movaps %xmm1, 0x2e0(%rsp)
movaps %xmm1, 0x2d0(%rsp)
movaps 0x2e0(%rsp), %xmm1
movaps 0x2d0(%rsp), %xmm2
mulps %xmm2, %xmm1
movaps %xmm1, 0xa0(%rsp)
movaps 0x7d8a70(%rip), %xmm1 # 0x1e1ca70
movaps %xmm1, 0x110(%rsp)
movaps 0x110(%rsp), %xmm2
movaps 0xa0(%rsp), %xmm1
movaps %xmm2, 0x5c0(%rsp)
movaps %xmm1, 0x5b0(%rsp)
movaps 0x7d8a51(%rip), %xmm1 # 0x1e1ca80
movaps %xmm1, 0x5a0(%rsp)
movaps 0x5c0(%rsp), %xmm2
movaps 0x5b0(%rsp), %xmm1
movaps %xmm2, 0x5e0(%rsp)
movaps %xmm1, 0x5d0(%rsp)
movaps 0x5e0(%rsp), %xmm2
movaps 0x5d0(%rsp), %xmm1
mulps %xmm1, %xmm2
movaps 0x5a0(%rsp), %xmm1
movaps %xmm2, 0x600(%rsp)
movaps %xmm1, 0x5f0(%rsp)
movaps 0x600(%rsp), %xmm1
movaps 0x5f0(%rsp), %xmm2
addps %xmm2, %xmm1
movaps %xmm1, 0x110(%rsp)
movaps 0x110(%rsp), %xmm2
movaps 0xa0(%rsp), %xmm1
movaps %xmm2, 0x550(%rsp)
movaps %xmm1, 0x540(%rsp)
movaps 0x7d89cc(%rip), %xmm1 # 0x1e1ca90
movaps %xmm1, 0x530(%rsp)
movaps 0x550(%rsp), %xmm2
movaps 0x540(%rsp), %xmm1
movaps %xmm2, 0x570(%rsp)
movaps %xmm1, 0x560(%rsp)
movaps 0x570(%rsp), %xmm2
movaps 0x560(%rsp), %xmm1
mulps %xmm1, %xmm2
movaps 0x530(%rsp), %xmm1
movaps %xmm2, 0x590(%rsp)
movaps %xmm1, 0x580(%rsp)
movaps 0x590(%rsp), %xmm1
movaps 0x580(%rsp), %xmm2
addps %xmm2, %xmm1
movaps %xmm1, 0x110(%rsp)
movaps 0x110(%rsp), %xmm2
movaps 0xa0(%rsp), %xmm1
movaps %xmm2, 0x2c0(%rsp)
movaps %xmm1, 0x2b0(%rsp)
movaps 0x2c0(%rsp), %xmm1
movaps 0x2b0(%rsp), %xmm2
mulps %xmm2, %xmm1
movaps %xmm1, 0x110(%rsp)
movaps 0x110(%rsp), %xmm2
movaps 0xa0(%rsp), %xmm1
movaps %xmm2, 0x2a0(%rsp)
movaps %xmm1, 0x290(%rsp)
movaps 0x2a0(%rsp), %xmm1
movaps 0x290(%rsp), %xmm2
mulps %xmm2, %xmm1
movaps %xmm1, 0x110(%rsp)
movaps 0xa0(%rsp), %xmm2
movaps 0x110(%rsp), %xmm1
movaps %xmm2, 0x390(%rsp)
movaps 0x7c78b9(%rip), %xmm2 # 0x1e0ba80
movaps %xmm2, 0x380(%rsp)
movaps %xmm1, 0x370(%rsp)
movaps 0x370(%rsp), %xmm2
movaps 0x390(%rsp), %xmm3
movaps 0x380(%rsp), %xmm1
movaps %xmm3, 0x3d0(%rsp)
movaps %xmm1, 0x3c0(%rsp)
movaps 0x3d0(%rsp), %xmm1
movaps 0x3c0(%rsp), %xmm3
mulps %xmm3, %xmm1
movaps %xmm2, 0x3b0(%rsp)
movaps %xmm1, 0x3a0(%rsp)
movaps 0x3b0(%rsp), %xmm1
movaps 0x3a0(%rsp), %xmm2
subps %xmm2, %xmm1
movaps %xmm1, 0x110(%rsp)
movaps 0x110(%rsp), %xmm1
movaps %xmm1, 0x360(%rsp)
movaps 0x7c77ec(%rip), %xmm1 # 0x1e0ba40
movaps %xmm1, 0x350(%rsp)
movaps 0x360(%rsp), %xmm1
movaps 0x350(%rsp), %xmm2
addps %xmm2, %xmm1
movaps %xmm1, 0x110(%rsp)
movaps 0x7d8822(%rip), %xmm1 # 0x1e1caa0
movaps %xmm1, 0x90(%rsp)
movaps 0x90(%rsp), %xmm2
movaps 0xa0(%rsp), %xmm1
movaps %xmm2, 0x4e0(%rsp)
movaps %xmm1, 0x4d0(%rsp)
movaps 0x7d8803(%rip), %xmm1 # 0x1e1cab0
movaps %xmm1, 0x4c0(%rsp)
movaps 0x4e0(%rsp), %xmm2
movaps 0x4d0(%rsp), %xmm1
movaps %xmm2, 0x500(%rsp)
movaps %xmm1, 0x4f0(%rsp)
movaps 0x500(%rsp), %xmm2
movaps 0x4f0(%rsp), %xmm1
mulps %xmm1, %xmm2
movaps 0x4c0(%rsp), %xmm1
movaps %xmm2, 0x520(%rsp)
movaps %xmm1, 0x510(%rsp)
movaps 0x520(%rsp), %xmm1
movaps 0x510(%rsp), %xmm2
addps %xmm2, %xmm1
movaps %xmm1, 0x90(%rsp)
movaps 0x90(%rsp), %xmm2
movaps 0xa0(%rsp), %xmm1
movaps %xmm2, 0x470(%rsp)
movaps %xmm1, 0x460(%rsp)
movaps 0x7d877e(%rip), %xmm1 # 0x1e1cac0
movaps %xmm1, 0x450(%rsp)
movaps 0x470(%rsp), %xmm2
movaps 0x460(%rsp), %xmm1
movaps %xmm2, 0x490(%rsp)
movaps %xmm1, 0x480(%rsp)
movaps 0x490(%rsp), %xmm2
movaps 0x480(%rsp), %xmm1
mulps %xmm1, %xmm2
movaps 0x450(%rsp), %xmm1
movaps %xmm2, 0x4b0(%rsp)
movaps %xmm1, 0x4a0(%rsp)
movaps 0x4b0(%rsp), %xmm1
movaps 0x4a0(%rsp), %xmm2
addps %xmm2, %xmm1
movaps %xmm1, 0x90(%rsp)
movaps 0x90(%rsp), %xmm2
movaps 0xa0(%rsp), %xmm1
movaps %xmm2, 0x280(%rsp)
movaps %xmm1, 0x270(%rsp)
movaps 0x280(%rsp), %xmm1
movaps 0x270(%rsp), %xmm2
mulps %xmm2, %xmm1
movaps %xmm1, 0x90(%rsp)
movaps 0x90(%rsp), %xmm2
movaps 0x170(%rsp), %xmm1
movaps %xmm2, 0x400(%rsp)
movaps %xmm1, 0x3f0(%rsp)
movaps %xmm1, 0x3e0(%rsp)
movaps 0x400(%rsp), %xmm2
movaps 0x3f0(%rsp), %xmm1
movaps %xmm2, 0x420(%rsp)
movaps %xmm1, 0x410(%rsp)
movaps 0x420(%rsp), %xmm2
movaps 0x410(%rsp), %xmm1
mulps %xmm1, %xmm2
movaps 0x3e0(%rsp), %xmm1
movaps %xmm2, 0x440(%rsp)
movaps %xmm1, 0x430(%rsp)
movaps 0x440(%rsp), %xmm1
movaps 0x430(%rsp), %xmm2
addps %xmm2, %xmm1
movaps %xmm1, 0x90(%rsp)
movaps 0xc0(%rsp), %xmm1
movaps %xmm1, 0x130(%rsp)
movaps 0x130(%rsp), %xmm2
movaps 0x90(%rsp), %xmm1
movaps %xmm2, 0x200(%rsp)
movaps %xmm1, 0x1f0(%rsp)
movaps 0x200(%rsp), %xmm1
movaps 0x1f0(%rsp), %xmm2
pand %xmm2, %xmm1
movaps %xmm1, 0x80(%rsp)
movaps 0x130(%rsp), %xmm2
movaps 0x110(%rsp), %xmm1
movaps %xmm2, 0x190(%rsp)
movaps %xmm1, 0x180(%rsp)
movaps 0x190(%rsp), %xmm1
movaps 0x180(%rsp), %xmm2
pandn %xmm2, %xmm1
movaps %xmm1, 0x70(%rsp)
movaps 0x90(%rsp), %xmm2
movaps 0x80(%rsp), %xmm1
movaps %xmm2, 0x1d0(%rsp)
movaps %xmm1, 0x1c0(%rsp)
movaps 0x1d0(%rsp), %xmm1
movaps 0x1c0(%rsp), %xmm2
subps %xmm2, %xmm1
movaps %xmm1, 0x90(%rsp)
movaps 0x110(%rsp), %xmm2
movaps 0x70(%rsp), %xmm1
movaps %xmm2, 0x1b0(%rsp)
movaps %xmm1, 0x1a0(%rsp)
movaps 0x1b0(%rsp), %xmm1
movaps 0x1a0(%rsp), %xmm2
subps %xmm2, %xmm1
movaps %xmm1, 0x110(%rsp)
movaps 0x70(%rsp), %xmm2
movaps 0x80(%rsp), %xmm1
movaps %xmm2, 0x340(%rsp)
movaps %xmm1, 0x330(%rsp)
movaps 0x340(%rsp), %xmm1
movaps 0x330(%rsp), %xmm2
addps %xmm2, %xmm1
movaps %xmm1, 0x150(%rsp)
movaps 0x110(%rsp), %xmm2
movaps 0x90(%rsp), %xmm1
movaps %xmm2, 0x320(%rsp)
movaps %xmm1, 0x310(%rsp)
movaps 0x320(%rsp), %xmm1
movaps 0x310(%rsp), %xmm2
addps %xmm2, %xmm1
movaps %xmm1, 0x140(%rsp)
movaps 0x150(%rsp), %xmm2
movaps 0x120(%rsp), %xmm1
movaps %xmm2, 0x8d0(%rsp)
movaps %xmm1, 0x8c0(%rsp)
movaps 0x8d0(%rsp), %xmm1
movaps 0x8c0(%rsp), %xmm2
pxor %xmm2, %xmm1
movq 0x168(%rsp), %rax
movaps %xmm1, (%rax)
movaps 0x140(%rsp), %xmm2
movaps 0xb0(%rsp), %xmm1
movaps %xmm2, 0x8b0(%rsp)
movaps %xmm1, 0x8a0(%rsp)
movaps 0x8b0(%rsp), %xmm1
movaps 0x8a0(%rsp), %xmm2
pxor %xmm2, %xmm1
movq 0x160(%rsp), %rax
movaps %xmm1, (%rax)
movaps -0x30(%rsp), %xmm1
movaps %xmm0, 0x20(%rsp)
movaps 0x20(%rsp), %xmm0
movaps %xmm1, 0x930(%rsp)
movaps %xmm0, 0x920(%rsp)
movaps 0x930(%rsp), %xmm0
movaps 0x920(%rsp), %xmm1
cmpeqps %xmm1, %xmm0
movaps %xmm0, -0x50(%rsp)
movaps -0x40(%rsp), %xmm1
movaps -0x50(%rsp), %xmm0
movaps %xmm1, 0x40(%rsp)
movaps %xmm0, 0x30(%rsp)
movaps 0x40(%rsp), %xmm0
movaps 0x30(%rsp), %xmm1
pand %xmm1, %xmm0
movaps %xmm0, -0x60(%rsp)
movaps -0x30(%rsp), %xmm1
movaps -0x60(%rsp), %xmm0
movaps %xmm1, 0x60(%rsp)
movaps %xmm0, 0x50(%rsp)
movaps 0x60(%rsp), %xmm0
addps 0x50(%rsp), %xmm0
movaps %xmm0, -0x30(%rsp)
movaps -0x20(%rsp), %xmm1
movaps -0x30(%rsp), %xmm0
movaps %xmm1, 0x950(%rsp)
movaps %xmm0, 0x940(%rsp)
movaps 0x950(%rsp), %xmm0
divps 0x940(%rsp), %xmm0
movaps %xmm0, -0x70(%rsp)
movaps -0x70(%rsp), %xmm0
addq $0x968, %rsp # imm = 0x968
retq
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,221 |
ncnn::UnaryOp_x86_functor::unary_op_tan::func(float const&) const
|
float func(const float& x) const
{
return (float)tan(x);
}
|
subq $0x18, %rsp
movq %rdi, 0x10(%rsp)
movq %rsi, 0x8(%rsp)
movq 0x8(%rsp), %rax
movss (%rax), %xmm0
callq 0x163a300
addq $0x18, %rsp
retq
nopw %cs:(%rax,%rax)
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,222 |
ncnn::UnaryOp_x86_functor::unary_op_asin::func_pack4(float vector[4] const&) const
|
__m128 func_pack4(const __m128& x) const
{
//TODO sse optimize
float tmp[4];
_mm_storeu_ps(tmp, x);
tmp[0] = asin(tmp[0]);
tmp[1] = asin(tmp[1]);
tmp[2] = asin(tmp[2]);
tmp[3] = asin(tmp[3]);
return _mm_loadu_ps(tmp);
}
|
subq $0x48, %rsp
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
movq %rsp, %rax
movq 0x10(%rsp), %rcx
movaps (%rcx), %xmm0
movq %rax, 0x38(%rsp)
movaps %xmm0, 0x20(%rsp)
movaps 0x20(%rsp), %xmm0
movq 0x38(%rsp), %rax
movups %xmm0, (%rax)
movss (%rsp), %xmm0
callq 0x163a350
movss %xmm0, (%rsp)
movss 0x4(%rsp), %xmm0
callq 0x163a350
movss %xmm0, 0x4(%rsp)
movss 0x8(%rsp), %xmm0
callq 0x163a350
movss %xmm0, 0x8(%rsp)
movss 0xc(%rsp), %xmm0
callq 0x163a350
movss %xmm0, 0xc(%rsp)
movq %rsp, %rax
movq %rax, 0x40(%rsp)
movq 0x40(%rsp), %rax
movups (%rax), %xmm0
addq $0x48, %rsp
retq
nopw (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,223 |
ncnn::UnaryOp_x86_functor::unary_op_asin::func(float const&) const
|
float func(const float& x) const
{
return (float)asin(x);
}
|
subq $0x18, %rsp
movq %rdi, 0x10(%rsp)
movq %rsi, 0x8(%rsp)
movq 0x8(%rsp), %rax
movss (%rax), %xmm0
callq 0x163a350
addq $0x18, %rsp
retq
nopw %cs:(%rax,%rax)
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,224 |
ncnn::UnaryOp_x86_functor::unary_op_acos::func_pack4(float vector[4] const&) const
|
__m128 func_pack4(const __m128& x) const
{
//TODO sse optimize
float tmp[4];
_mm_storeu_ps(tmp, x);
tmp[0] = acos(tmp[0]);
tmp[1] = acos(tmp[1]);
tmp[2] = acos(tmp[2]);
tmp[3] = acos(tmp[3]);
return _mm_loadu_ps(tmp);
}
|
subq $0x48, %rsp
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
movq %rsp, %rax
movq 0x10(%rsp), %rcx
movaps (%rcx), %xmm0
movq %rax, 0x38(%rsp)
movaps %xmm0, 0x20(%rsp)
movaps 0x20(%rsp), %xmm0
movq 0x38(%rsp), %rax
movups %xmm0, (%rax)
movss (%rsp), %xmm0
callq 0x163a3a0
movss %xmm0, (%rsp)
movss 0x4(%rsp), %xmm0
callq 0x163a3a0
movss %xmm0, 0x4(%rsp)
movss 0x8(%rsp), %xmm0
callq 0x163a3a0
movss %xmm0, 0x8(%rsp)
movss 0xc(%rsp), %xmm0
callq 0x163a3a0
movss %xmm0, 0xc(%rsp)
movq %rsp, %rax
movq %rax, 0x40(%rsp)
movq 0x40(%rsp), %rax
movups (%rax), %xmm0
addq $0x48, %rsp
retq
nopw (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,225 |
ncnn::UnaryOp_x86_functor::unary_op_acos::func(float const&) const
|
float func(const float& x) const
{
return (float)acos(x);
}
|
subq $0x18, %rsp
movq %rdi, 0x10(%rsp)
movq %rsi, 0x8(%rsp)
movq 0x8(%rsp), %rax
movss (%rax), %xmm0
callq 0x163a3a0
addq $0x18, %rsp
retq
nopw %cs:(%rax,%rax)
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,226 |
ncnn::UnaryOp_x86_functor::unary_op_atan::func_pack4(float vector[4] const&) const
|
__m128 func_pack4(const __m128& x) const
{
//TODO sse optimize
float tmp[4];
_mm_storeu_ps(tmp, x);
tmp[0] = atan(tmp[0]);
tmp[1] = atan(tmp[1]);
tmp[2] = atan(tmp[2]);
tmp[3] = atan(tmp[3]);
return _mm_loadu_ps(tmp);
}
|
subq $0x48, %rsp
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
movq %rsp, %rax
movq 0x10(%rsp), %rcx
movaps (%rcx), %xmm0
movq %rax, 0x38(%rsp)
movaps %xmm0, 0x20(%rsp)
movaps 0x20(%rsp), %xmm0
movq 0x38(%rsp), %rax
movups %xmm0, (%rax)
movss (%rsp), %xmm0
callq 0x163a3f0
movss %xmm0, (%rsp)
movss 0x4(%rsp), %xmm0
callq 0x163a3f0
movss %xmm0, 0x4(%rsp)
movss 0x8(%rsp), %xmm0
callq 0x163a3f0
movss %xmm0, 0x8(%rsp)
movss 0xc(%rsp), %xmm0
callq 0x163a3f0
movss %xmm0, 0xc(%rsp)
movq %rsp, %rax
movq %rax, 0x40(%rsp)
movq 0x40(%rsp), %rax
movups (%rax), %xmm0
addq $0x48, %rsp
retq
nopw (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,227 |
ncnn::UnaryOp_x86_functor::unary_op_atan::func(float const&) const
|
float func(const float& x) const
{
return (float)atan(x);
}
|
subq $0x18, %rsp
movq %rdi, 0x10(%rsp)
movq %rsi, 0x8(%rsp)
movq 0x8(%rsp), %rax
movss (%rax), %xmm0
callq 0x163a3f0
addq $0x18, %rsp
retq
nopw %cs:(%rax,%rax)
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,228 |
ncnn::UnaryOp_x86_functor::unary_op_reciprocal::func_pack4(float vector[4] const&) const
|
__m128 func_pack4(const __m128& x) const
{
return _mm_div_ps(*(__m128*)_ps_1, x);
}
|
movq %rdi, -0x30(%rsp)
movq %rsi, -0x38(%rsp)
movaps 0x7d814f(%rip), %xmm1 # 0x1e1caf0
movq -0x38(%rsp), %rax
movaps (%rax), %xmm0
movaps %xmm1, -0x18(%rsp)
movaps %xmm0, -0x28(%rsp)
movaps -0x18(%rsp), %xmm0
divps -0x28(%rsp), %xmm0
retq
nop
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,229 |
ncnn::UnaryOp_x86_functor::unary_op_reciprocal::func(float const&) const
|
float func(const float& x) const
{
return 1.f / x;
}
|
movq %rdi, -0x8(%rsp)
movq %rsi, -0x10(%rsp)
movq -0x10(%rsp), %rax
movss 0x7c462d(%rip), %xmm0 # 0x1e09004
divss (%rax), %xmm0
retq
nopl (%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,230 |
ncnn::UnaryOp_x86_functor::unary_op_tanh::func_pack4(float vector[4] const&) const
|
__m128 func_pack4(const __m128& x) const
{
return tanh_sse(x);
}
|
subq $0x6e8, %rsp # imm = 0x6E8
movq %rdi, -0x78(%rsp)
movq %rsi, -0x80(%rsp)
movq -0x80(%rsp), %rax
movaps (%rax), %xmm0
movaps %xmm0, -0x50(%rsp)
movl $0x3f800000, -0x4(%rsp) # imm = 0x3F800000
movss -0x4(%rsp), %xmm0
shufps $0x0, %xmm0, %xmm0 # xmm0 = xmm0[0,0,0,0]
movaps %xmm0, -0x20(%rsp)
movaps -0x20(%rsp), %xmm0
movaps %xmm0, -0x60(%rsp)
movl $0x40000000, -0x24(%rsp) # imm = 0x40000000
movss -0x24(%rsp), %xmm0
shufps $0x0, %xmm0, %xmm0 # xmm0 = xmm0[0,0,0,0]
movaps %xmm0, -0x40(%rsp)
movaps -0x40(%rsp), %xmm0
movaps %xmm0, -0x70(%rsp)
movaps -0x50(%rsp), %xmm1
movaps -0x70(%rsp), %xmm0
movaps %xmm1, 0x50(%rsp)
movaps %xmm0, 0x40(%rsp)
movaps 0x50(%rsp), %xmm0
movaps 0x40(%rsp), %xmm1
mulps %xmm1, %xmm0
movaps %xmm0, 0x70(%rsp)
movl $0x3f800000, 0x9c(%rsp) # imm = 0x3F800000
movss 0x9c(%rsp), %xmm0
shufps $0x0, %xmm0, %xmm0 # xmm0 = xmm0[0,0,0,0]
movaps %xmm0, 0x80(%rsp)
movaps 0x80(%rsp), %xmm0
movaps %xmm0, 0x60(%rsp)
movaps 0x60(%rsp), %xmm2
movaps %xmm2, %xmm1
xorps %xmm0, %xmm0
movaps %xmm0, 0xc0(%rsp)
movaps 0xc0(%rsp), %xmm4
movaps 0x70(%rsp), %xmm3
movaps %xmm4, 0xb0(%rsp)
movaps %xmm3, 0xa0(%rsp)
movaps 0xb0(%rsp), %xmm3
movaps 0xa0(%rsp), %xmm4
subps %xmm4, %xmm3
movaps %xmm3, 0x140(%rsp)
movaps %xmm0, 0x170(%rsp)
movaps 0x170(%rsp), %xmm0
movaps %xmm0, 0x130(%rsp)
movaps 0x7c6f43(%rip), %xmm0 # 0x1e0ba40
movaps %xmm0, 0x100(%rsp)
movaps 0x140(%rsp), %xmm0
movaps %xmm0, 0x260(%rsp)
movaps 0x7c6f34(%rip), %xmm0 # 0x1e0ba50
movaps %xmm0, 0x250(%rsp)
movaps 0x260(%rsp), %xmm0
movaps 0x250(%rsp), %xmm3
minps %xmm3, %xmm0
movaps %xmm0, 0x140(%rsp)
movaps 0x140(%rsp), %xmm0
movaps %xmm0, 0x280(%rsp)
movaps 0x7c6f0a(%rip), %xmm0 # 0x1e0ba60
movaps %xmm0, 0x270(%rsp)
movaps 0x280(%rsp), %xmm0
movaps 0x270(%rsp), %xmm3
maxps %xmm3, %xmm0
movaps %xmm0, 0x140(%rsp)
movaps 0x140(%rsp), %xmm0
movaps %xmm0, 0x240(%rsp)
movaps 0x7c6ee0(%rip), %xmm0 # 0x1e0ba70
movaps %xmm0, 0x230(%rsp)
movaps 0x240(%rsp), %xmm0
movaps 0x230(%rsp), %xmm3
mulps %xmm3, %xmm0
movaps %xmm0, 0x120(%rsp)
movaps 0x120(%rsp), %xmm0
movaps %xmm0, 0x2c0(%rsp)
movaps 0x7c6eb6(%rip), %xmm0 # 0x1e0ba80
movaps %xmm0, 0x2b0(%rsp)
movaps 0x2c0(%rsp), %xmm3
movaps 0x2b0(%rsp), %xmm4
addps %xmm4, %xmm3
movaps %xmm3, 0x120(%rsp)
movaps 0x120(%rsp), %xmm3
movaps %xmm3, 0x1c0(%rsp)
cvttps2dq 0x1c0(%rsp), %xmm3
movaps %xmm3, 0x110(%rsp)
movaps 0x110(%rsp), %xmm3
movaps %xmm3, 0x1a0(%rsp)
cvtdq2ps 0x1a0(%rsp), %xmm3
movaps %xmm3, 0x130(%rsp)
movaps 0x130(%rsp), %xmm4
movaps 0x120(%rsp), %xmm3
movaps %xmm4, 0x1e0(%rsp)
movaps %xmm3, 0x1d0(%rsp)
movaps 0x1d0(%rsp), %xmm3
movaps 0x1e0(%rsp), %xmm4
cmpltps %xmm4, %xmm3
movaps %xmm3, 0xf0(%rsp)
movaps 0xf0(%rsp), %xmm4
movaps 0x100(%rsp), %xmm3
movaps %xmm4, 0x190(%rsp)
movaps %xmm3, 0x180(%rsp)
movaps 0x190(%rsp), %xmm3
movaps 0x180(%rsp), %xmm4
pand %xmm4, %xmm3
movaps %xmm3, 0xf0(%rsp)
movaps 0x130(%rsp), %xmm4
movaps 0xf0(%rsp), %xmm3
movaps %xmm4, 0x160(%rsp)
movaps %xmm3, 0x150(%rsp)
movaps 0x160(%rsp), %xmm3
movaps 0x150(%rsp), %xmm4
subps %xmm4, %xmm3
movaps %xmm3, 0x120(%rsp)
movaps 0x120(%rsp), %xmm4
movaps 0x140(%rsp), %xmm3
movaps %xmm4, 0x380(%rsp)
movaps 0x7c6d90(%rip), %xmm4 # 0x1e0ba90
movaps %xmm4, 0x370(%rsp)
movaps %xmm3, 0x360(%rsp)
movaps 0x360(%rsp), %xmm4
movaps 0x380(%rsp), %xmm5
movaps 0x370(%rsp), %xmm3
movaps %xmm5, 0x3c0(%rsp)
movaps %xmm3, 0x3b0(%rsp)
movaps 0x3c0(%rsp), %xmm3
movaps 0x3b0(%rsp), %xmm5
mulps %xmm5, %xmm3
movaps %xmm4, 0x3a0(%rsp)
movaps %xmm3, 0x390(%rsp)
movaps 0x3a0(%rsp), %xmm3
movaps 0x390(%rsp), %xmm4
subps %xmm4, %xmm3
movaps %xmm3, 0x140(%rsp)
movaps 0x120(%rsp), %xmm4
movaps 0x140(%rsp), %xmm3
movaps %xmm4, 0x310(%rsp)
movaps 0x7c6d0b(%rip), %xmm4 # 0x1e0baa0
movaps %xmm4, 0x300(%rsp)
movaps %xmm3, 0x2f0(%rsp)
movaps 0x2f0(%rsp), %xmm4
movaps 0x310(%rsp), %xmm5
movaps 0x300(%rsp), %xmm3
movaps %xmm5, 0x350(%rsp)
movaps %xmm3, 0x340(%rsp)
movaps 0x350(%rsp), %xmm3
movaps 0x340(%rsp), %xmm5
mulps %xmm5, %xmm3
movaps %xmm4, 0x330(%rsp)
movaps %xmm3, 0x320(%rsp)
movaps 0x330(%rsp), %xmm3
movaps 0x320(%rsp), %xmm4
subps %xmm4, %xmm3
movaps %xmm3, 0x140(%rsp)
movaps 0x140(%rsp), %xmm3
movaps %xmm3, 0x220(%rsp)
movaps %xmm3, 0x210(%rsp)
movaps 0x220(%rsp), %xmm3
movaps 0x210(%rsp), %xmm4
mulps %xmm4, %xmm3
movaps %xmm3, 0x130(%rsp)
movaps 0x7c6c6b(%rip), %xmm3 # 0x1e0bab0
movaps %xmm3, 0xe0(%rsp)
movaps 0xe0(%rsp), %xmm4
movaps 0x140(%rsp), %xmm3
movaps %xmm4, 0x620(%rsp)
movaps %xmm3, 0x610(%rsp)
movaps 0x7c6c4c(%rip), %xmm3 # 0x1e0bac0
movaps %xmm3, 0x600(%rsp)
movaps 0x620(%rsp), %xmm4
movaps 0x610(%rsp), %xmm3
movaps %xmm4, 0x640(%rsp)
movaps %xmm3, 0x630(%rsp)
movaps 0x640(%rsp), %xmm4
movaps 0x630(%rsp), %xmm3
mulps %xmm3, %xmm4
movaps 0x600(%rsp), %xmm3
movaps %xmm4, 0x660(%rsp)
movaps %xmm3, 0x650(%rsp)
movaps 0x660(%rsp), %xmm3
movaps 0x650(%rsp), %xmm4
addps %xmm4, %xmm3
movaps %xmm3, 0xe0(%rsp)
movaps 0xe0(%rsp), %xmm4
movaps 0x140(%rsp), %xmm3
movaps %xmm4, 0x5b0(%rsp)
movaps %xmm3, 0x5a0(%rsp)
movaps 0x7c6bc7(%rip), %xmm3 # 0x1e0bad0
movaps %xmm3, 0x590(%rsp)
movaps 0x5b0(%rsp), %xmm4
movaps 0x5a0(%rsp), %xmm3
movaps %xmm4, 0x5d0(%rsp)
movaps %xmm3, 0x5c0(%rsp)
movaps 0x5d0(%rsp), %xmm4
movaps 0x5c0(%rsp), %xmm3
mulps %xmm3, %xmm4
movaps 0x590(%rsp), %xmm3
movaps %xmm4, 0x5f0(%rsp)
movaps %xmm3, 0x5e0(%rsp)
movaps 0x5f0(%rsp), %xmm3
movaps 0x5e0(%rsp), %xmm4
addps %xmm4, %xmm3
movaps %xmm3, 0xe0(%rsp)
movaps 0xe0(%rsp), %xmm4
movaps 0x140(%rsp), %xmm3
movaps %xmm4, 0x540(%rsp)
movaps %xmm3, 0x530(%rsp)
movaps 0x7c6b42(%rip), %xmm3 # 0x1e0bae0
movaps %xmm3, 0x520(%rsp)
movaps 0x540(%rsp), %xmm4
movaps 0x530(%rsp), %xmm3
movaps %xmm4, 0x560(%rsp)
movaps %xmm3, 0x550(%rsp)
movaps 0x560(%rsp), %xmm4
movaps 0x550(%rsp), %xmm3
mulps %xmm3, %xmm4
movaps 0x520(%rsp), %xmm3
movaps %xmm4, 0x580(%rsp)
movaps %xmm3, 0x570(%rsp)
movaps 0x580(%rsp), %xmm3
movaps 0x570(%rsp), %xmm4
addps %xmm4, %xmm3
movaps %xmm3, 0xe0(%rsp)
movaps 0xe0(%rsp), %xmm4
movaps 0x140(%rsp), %xmm3
movaps %xmm4, 0x4d0(%rsp)
movaps %xmm3, 0x4c0(%rsp)
movaps 0x7c6abd(%rip), %xmm3 # 0x1e0baf0
movaps %xmm3, 0x4b0(%rsp)
movaps 0x4d0(%rsp), %xmm4
movaps 0x4c0(%rsp), %xmm3
movaps %xmm4, 0x4f0(%rsp)
movaps %xmm3, 0x4e0(%rsp)
movaps 0x4f0(%rsp), %xmm4
movaps 0x4e0(%rsp), %xmm3
mulps %xmm3, %xmm4
movaps 0x4b0(%rsp), %xmm3
movaps %xmm4, 0x510(%rsp)
movaps %xmm3, 0x500(%rsp)
movaps 0x510(%rsp), %xmm3
movaps 0x500(%rsp), %xmm4
addps %xmm4, %xmm3
movaps %xmm3, 0xe0(%rsp)
movaps 0xe0(%rsp), %xmm4
movaps 0x140(%rsp), %xmm3
movaps %xmm4, 0x460(%rsp)
movaps %xmm3, 0x450(%rsp)
movaps %xmm0, 0x440(%rsp)
movaps 0x460(%rsp), %xmm3
movaps 0x450(%rsp), %xmm0
movaps %xmm3, 0x480(%rsp)
movaps %xmm0, 0x470(%rsp)
movaps 0x480(%rsp), %xmm3
movaps 0x470(%rsp), %xmm0
mulps %xmm0, %xmm3
movaps 0x440(%rsp), %xmm0
movaps %xmm3, 0x4a0(%rsp)
movaps %xmm0, 0x490(%rsp)
movaps 0x4a0(%rsp), %xmm0
movaps 0x490(%rsp), %xmm3
addps %xmm3, %xmm0
movaps %xmm0, 0xe0(%rsp)
movaps 0xe0(%rsp), %xmm4
movaps 0x130(%rsp), %xmm3
movaps 0x140(%rsp), %xmm0
movaps %xmm4, 0x3f0(%rsp)
movaps %xmm3, 0x3e0(%rsp)
movaps %xmm0, 0x3d0(%rsp)
movaps 0x3f0(%rsp), %xmm3
movaps 0x3e0(%rsp), %xmm0
movaps %xmm3, 0x410(%rsp)
movaps %xmm0, 0x400(%rsp)
movaps 0x410(%rsp), %xmm3
movaps 0x400(%rsp), %xmm0
mulps %xmm0, %xmm3
movaps 0x3d0(%rsp), %xmm0
movaps %xmm3, 0x430(%rsp)
movaps %xmm0, 0x420(%rsp)
movaps 0x430(%rsp), %xmm0
movaps 0x420(%rsp), %xmm3
addps %xmm3, %xmm0
movaps %xmm0, 0xe0(%rsp)
movaps 0xe0(%rsp), %xmm3
movaps 0x100(%rsp), %xmm0
movaps %xmm3, 0x2a0(%rsp)
movaps %xmm0, 0x290(%rsp)
movaps 0x2a0(%rsp), %xmm0
movaps 0x290(%rsp), %xmm3
addps %xmm3, %xmm0
movaps %xmm0, 0xe0(%rsp)
movaps 0x120(%rsp), %xmm0
movaps %xmm0, 0x1b0(%rsp)
cvttps2dq 0x1b0(%rsp), %xmm0
movaps %xmm0, 0x110(%rsp)
movaps 0x110(%rsp), %xmm0
movaps %xmm0, 0x680(%rsp)
movaps 0x7c68c8(%rip), %xmm0 # 0x1e0bb00
movaps %xmm0, 0x670(%rsp)
movdqa 0x680(%rsp), %xmm0
movdqa 0x670(%rsp), %xmm3
paddd %xmm3, %xmm0
movdqa %xmm0, 0x110(%rsp)
movdqa 0x110(%rsp), %xmm0
movdqa %xmm0, 0x6a0(%rsp)
movl $0x17, 0x69c(%rsp)
movdqa 0x6a0(%rsp), %xmm0
movl 0x69c(%rsp), %eax
movd %eax, %xmm3
pslld %xmm3, %xmm0
movdqa %xmm0, 0x110(%rsp)
movdqa 0x110(%rsp), %xmm0
movdqa %xmm0, 0x6b0(%rsp)
movdqa 0x6b0(%rsp), %xmm0
movaps %xmm0, 0xd0(%rsp)
movaps 0xe0(%rsp), %xmm3
movaps 0xd0(%rsp), %xmm0
movaps %xmm3, 0x200(%rsp)
movaps %xmm0, 0x1f0(%rsp)
movaps 0x200(%rsp), %xmm0
mulps 0x1f0(%rsp), %xmm0
movaps %xmm0, 0xe0(%rsp)
movaps 0xe0(%rsp), %xmm0
movaps %xmm2, 0x2e0(%rsp)
movaps %xmm0, 0x2d0(%rsp)
movaps 0x2e0(%rsp), %xmm0
addps 0x2d0(%rsp), %xmm0
movaps %xmm1, 0x6d0(%rsp)
movaps %xmm0, 0x6c0(%rsp)
movaps 0x6d0(%rsp), %xmm1
divps 0x6c0(%rsp), %xmm1
movaps -0x70(%rsp), %xmm0
movaps %xmm1, 0x30(%rsp)
movaps %xmm0, 0x20(%rsp)
movaps 0x30(%rsp), %xmm1
mulps 0x20(%rsp), %xmm1
movaps -0x60(%rsp), %xmm0
movaps %xmm1, 0x10(%rsp)
movaps %xmm0, (%rsp)
movaps 0x10(%rsp), %xmm0
subps (%rsp), %xmm0
addq $0x6e8, %rsp # imm = 0x6E8
retq
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,231 |
ncnn::UnaryOp_x86_functor::unary_op_tanh::func(float const&) const
|
float func(const float& x) const
{
return (float)tanh(x);
}
|
subq $0x18, %rsp
movq %rdi, 0x10(%rsp)
movq %rsi, 0x8(%rsp)
movq 0x8(%rsp), %rax
movss (%rax), %xmm0
callq 0x140cc0
addq $0x18, %rsp
retq
nopw %cs:(%rax,%rax)
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,232 |
ncnn::UnaryOp_x86_avx512::UnaryOp_x86_avx512()
|
UnaryOp_x86_avx512::UnaryOp_x86_avx512()
{
#if __SSE2__
support_packing = true;
#endif // __SSE2__
}
|
movq %rdi, -0x8(%rsp)
movq %rsi, -0x10(%rsp)
movq -0x8(%rsp), %rax
movq -0x10(%rsp), %rcx
movq (%rcx), %rdx
movq %rdx, (%rax)
movq 0x8(%rcx), %rdx
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
movq %rdx, (%rax,%rcx)
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
movb $0x1, 0xb(%rax,%rcx)
retq
nopw %cs:(%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,233 |
ncnn::UnaryOp_x86_avx512::UnaryOp_x86_avx512()
|
UnaryOp_x86_avx512::UnaryOp_x86_avx512()
{
#if __SSE2__
support_packing = true;
#endif // __SSE2__
}
|
subq $0x18, %rsp
movq %rdi, 0x10(%rsp)
movq 0x10(%rsp), %rdi
movq %rdi, 0x8(%rsp)
addq $0x8, %rdi
callq 0x1639190
movq 0x8(%rsp), %rax
leaq 0x88c760(%rip), %rcx # 0x1ed1b78
addq $0x18, %rcx
movq %rcx, (%rax)
leaq 0x88c752(%rip), %rcx # 0x1ed1b78
addq $0x88, %rcx
movq %rcx, 0x8(%rax)
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
movb $0x1, 0xb(%rax,%rcx)
addq $0x18, %rsp
retq
nopw %cs:(%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,234 |
ncnn::UnaryOp_x86_avx512::forward_inplace(ncnn::Mat&, ncnn::Option const&) const
|
int UnaryOp_x86_avx512::forward_inplace(Mat& bottom_top_blob, const Option& opt) const
{
using namespace UnaryOp_x86_avx512_functor;
if (op_type == Operation_ABS)
return unary_op_inplace<unary_op_abs>(bottom_top_blob, opt);
if (op_type == Operation_NEG)
return unary_op_inplace<unary_op_neg>(bottom_top_blob, opt);
if (op_type == Operation_FLOOR)
return unary_op_inplace<unary_op_floor>(bottom_top_blob, opt);
if (op_type == Operation_CEIL)
return unary_op_inplace<unary_op_ceil>(bottom_top_blob, opt);
if (op_type == Operation_SQUARE)
return unary_op_inplace<unary_op_square>(bottom_top_blob, opt);
if (op_type == Operation_SQRT)
return unary_op_inplace<unary_op_sqrt>(bottom_top_blob, opt);
if (op_type == Operation_RSQRT)
return unary_op_inplace<unary_op_rsqrt>(bottom_top_blob, opt);
if (op_type == Operation_EXP)
return unary_op_inplace<unary_op_exp>(bottom_top_blob, opt);
if (op_type == Operation_LOG)
return unary_op_inplace<unary_op_log>(bottom_top_blob, opt);
if (op_type == Operation_SIN)
return unary_op_inplace<unary_op_sin>(bottom_top_blob, opt);
if (op_type == Operation_COS)
return unary_op_inplace<unary_op_cos>(bottom_top_blob, opt);
if (op_type == Operation_TAN)
return unary_op_inplace<unary_op_tan>(bottom_top_blob, opt);
if (op_type == Operation_ASIN)
return unary_op_inplace<unary_op_asin>(bottom_top_blob, opt);
if (op_type == Operation_ACOS)
return unary_op_inplace<unary_op_acos>(bottom_top_blob, opt);
if (op_type == Operation_ATAN)
return unary_op_inplace<unary_op_atan>(bottom_top_blob, opt);
if (op_type == Operation_RECIPROCAL)
return unary_op_inplace<unary_op_reciprocal>(bottom_top_blob, opt);
if (op_type == Operation_TANH)
return unary_op_inplace<unary_op_tanh>(bottom_top_blob, opt);
return 0;
}
|
subq $0x28, %rsp
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
movq %rdx, 0x8(%rsp)
movq 0x18(%rsp), %rax
movq %rax, (%rsp)
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
cmpl $0x0, 0xd0(%rax,%rcx)
jne 0x1645495
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x1645770
movl %eax, 0x24(%rsp)
jmp 0x1645764
movq (%rsp), %rax
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
cmpl $0x1, 0xd0(%rax,%rcx)
jne 0x16454c2
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x1645f00
movl %eax, 0x24(%rsp)
jmp 0x1645764
movq (%rsp), %rax
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
cmpl $0x2, 0xd0(%rax,%rcx)
jne 0x16454ef
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x1646690
movl %eax, 0x24(%rsp)
jmp 0x1645764
movq (%rsp), %rax
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
cmpl $0x3, 0xd0(%rax,%rcx)
jne 0x164551c
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x1646e20
movl %eax, 0x24(%rsp)
jmp 0x1645764
movq (%rsp), %rax
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
cmpl $0x4, 0xd0(%rax,%rcx)
jne 0x1645549
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x16475b0
movl %eax, 0x24(%rsp)
jmp 0x1645764
movq (%rsp), %rax
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
cmpl $0x5, 0xd0(%rax,%rcx)
jne 0x1645576
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x1647d40
movl %eax, 0x24(%rsp)
jmp 0x1645764
movq (%rsp), %rax
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
cmpl $0x6, 0xd0(%rax,%rcx)
jne 0x16455a3
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x16484d0
movl %eax, 0x24(%rsp)
jmp 0x1645764
movq (%rsp), %rax
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
cmpl $0x7, 0xd0(%rax,%rcx)
jne 0x16455d0
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x1648c60
movl %eax, 0x24(%rsp)
jmp 0x1645764
movq (%rsp), %rax
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
cmpl $0x8, 0xd0(%rax,%rcx)
jne 0x16455fd
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x16493f0
movl %eax, 0x24(%rsp)
jmp 0x1645764
movq (%rsp), %rax
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
cmpl $0x9, 0xd0(%rax,%rcx)
jne 0x164562a
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x1649b80
movl %eax, 0x24(%rsp)
jmp 0x1645764
movq (%rsp), %rax
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
cmpl $0xa, 0xd0(%rax,%rcx)
jne 0x1645657
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x164a310
movl %eax, 0x24(%rsp)
jmp 0x1645764
movq (%rsp), %rax
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
cmpl $0xb, 0xd0(%rax,%rcx)
jne 0x1645684
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x164aaa0
movl %eax, 0x24(%rsp)
jmp 0x1645764
movq (%rsp), %rax
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
cmpl $0xc, 0xd0(%rax,%rcx)
jne 0x16456b1
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x164b230
movl %eax, 0x24(%rsp)
jmp 0x1645764
movq (%rsp), %rax
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
cmpl $0xd, 0xd0(%rax,%rcx)
jne 0x16456de
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x164b9c0
movl %eax, 0x24(%rsp)
jmp 0x1645764
movq (%rsp), %rax
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
cmpl $0xe, 0xd0(%rax,%rcx)
jne 0x1645708
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x164c150
movl %eax, 0x24(%rsp)
jmp 0x1645764
movq (%rsp), %rax
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
cmpl $0xf, 0xd0(%rax,%rcx)
jne 0x1645732
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x164c8e0
movl %eax, 0x24(%rsp)
jmp 0x1645764
movq (%rsp), %rax
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
cmpl $0x10, 0xd0(%rax,%rcx)
jne 0x164575c
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x164d070
movl %eax, 0x24(%rsp)
jmp 0x1645764
movl $0x0, 0x24(%rsp)
movl 0x24(%rsp), %eax
addq $0x28, %rsp
retq
nopl (%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,235 |
int ncnn::unary_op_inplace<ncnn::UnaryOp_x86_avx512_functor::unary_op_abs>(ncnn::Mat&, ncnn::Option const&)
|
static int unary_op_inplace(Mat& a, const Option& opt)
{
Op op;
int w = a.w;
int h = a.h;
int d = a.d;
int channels = a.c;
int elempack = a.elempack;
int size = w * h * d * elempack;
#pragma omp parallel for num_threads(opt.num_threads)
for (int q = 0; q < channels; q++)
{
float* ptr = a.channel(q);
int i = 0;
#if __SSE2__
#if __AVX__
#if __AVX512F__
for (; i + 15 < size; i += 16)
{
__m512 _p = _mm512_loadu_ps(ptr);
_p = op.func_pack16(_p);
_mm512_storeu_ps(ptr, _p);
ptr += 16;
}
#endif // __AVX512F__
for (; i + 7 < size; i += 8)
{
__m256 _p = _mm256_loadu_ps(ptr);
_p = op.func_pack8(_p);
_mm256_storeu_ps(ptr, _p);
ptr += 8;
}
#endif // __AVX__
for (; i + 3 < size; i += 4)
{
__m128 _p = _mm_load_ps(ptr);
_p = op.func_pack4(_p);
_mm_store_ps(ptr, _p);
ptr += 4;
}
#endif // __SSE2__
for (; i < size; i++)
{
*ptr = op.func(*ptr);
ptr++;
}
}
return 0;
}
|
pushq %rbp
movq %rsp, %rbp
andq $-0x40, %rsp
subq $0x340, %rsp # imm = 0x340
movq %rdi, 0x178(%rsp)
movq %rsi, 0x170(%rsp)
movq 0x178(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x168(%rsp)
movq 0x178(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0x164(%rsp)
movq 0x178(%rsp), %rax
movl 0x34(%rax), %eax
movl %eax, 0x160(%rsp)
movq 0x178(%rsp), %rax
movl 0x38(%rax), %eax
movl %eax, 0x15c(%rsp)
movq 0x178(%rsp), %rax
movl 0x18(%rax), %eax
movl %eax, 0x158(%rsp)
movl 0x168(%rsp), %eax
imull 0x164(%rsp), %eax
imull 0x160(%rsp), %eax
imull 0x158(%rsp), %eax
movl %eax, 0x154(%rsp)
movl $0x0, 0x150(%rsp)
movl 0x150(%rsp), %eax
cmpl 0x15c(%rsp), %eax
jge 0x1645ef0
movq 0x178(%rsp), %rcx
movl 0x150(%rsp), %eax
leaq 0x100(%rsp), %rdx
movq %rdx, 0x190(%rsp)
movq %rcx, 0x188(%rsp)
movl %eax, 0x184(%rsp)
movq 0x188(%rsp), %rax
movq %rax, 0x48(%rsp)
movb $0x0, 0x183(%rsp)
movl 0x2c(%rax), %r9d
movl 0x30(%rax), %r8d
movl 0x34(%rax), %edi
movq (%rax), %rsi
movq 0x40(%rax), %rcx
movslq 0x184(%rsp), %rdx
imulq %rdx, %rcx
imulq 0x10(%rax), %rcx
addq %rcx, %rsi
movq 0x10(%rax), %rdx
movl 0x18(%rax), %ecx
movq 0x20(%rax), %rax
leaq 0x100(%rsp), %r10
movq %r10, 0x2d0(%rsp)
movl %r9d, 0x2cc(%rsp)
movl %r8d, 0x2c8(%rsp)
movl %edi, 0x2c4(%rsp)
movq %rsi, 0x2b8(%rsp)
movq %rdx, 0x2b0(%rsp)
movl %ecx, 0x2ac(%rsp)
movq %rax, 0x2a0(%rsp)
movq 0x2d0(%rsp), %rcx
movq %rcx, 0x40(%rsp)
movq 0x2b8(%rsp), %rax
movq %rax, (%rcx)
movq $0x0, 0x8(%rcx)
movq 0x2b0(%rsp), %rax
movq %rax, 0x10(%rcx)
movl 0x2ac(%rsp), %eax
movl %eax, 0x18(%rcx)
movq 0x2a0(%rsp), %rax
movq %rax, 0x20(%rcx)
movl $0x3, 0x28(%rcx)
movl 0x2cc(%rsp), %eax
movl %eax, 0x2c(%rcx)
movl 0x2c8(%rsp), %eax
movl %eax, 0x30(%rcx)
movl $0x1, 0x34(%rcx)
movl 0x2c4(%rsp), %eax
movl %eax, 0x38(%rcx)
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rdx
imulq %rdx, %rax
imulq 0x10(%rcx), %rax
movq %rax, 0x2e0(%rsp)
movl $0x10, 0x2dc(%rsp)
movq 0x2e0(%rsp), %rax
movslq 0x2dc(%rsp), %rdx
addq %rdx, %rax
subq $0x1, %rax
xorl %edx, %edx
subl 0x2dc(%rsp), %edx
movslq %edx, %rdx
andq %rdx, %rax
xorl %edx, %edx
divq 0x10(%rcx)
movq %rax, %rdx
movq 0x48(%rsp), %rax
movq %rdx, 0x40(%rcx)
movl 0x28(%rax), %ecx
subl $0x1, %ecx
movl %ecx, 0x128(%rsp)
cmpl $0x4, 0x28(%rax)
jne 0x16459de
movq 0x48(%rsp), %rcx
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rcx
imulq %rcx, %rax
movq %rax, 0x140(%rsp)
movb $0x1, 0x183(%rsp)
testb $0x1, 0x183(%rsp)
jne 0x1645b0d
leaq 0x100(%rsp), %rax
movq %rax, 0x1a0(%rsp)
movq 0x1a0(%rsp), %rax
movq %rax, 0x310(%rsp)
movq 0x310(%rsp), %rax
movq %rax, 0x38(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x1645ab3
movq 0x38(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x30c(%rsp) # imm = 0xFFFFFFFF
movl 0x30c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x308(%rsp)
cmpl $0x1, 0x308(%rsp)
jne 0x1645ab3
movq 0x38(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x1645a84
movq 0x38(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x1645a82
jmp 0x1645ab1
movq 0x38(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x318(%rsp)
cmpq $0x0, 0x318(%rsp)
je 0x1645aaf
movq 0x318(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x1645ab1
jmp 0x1645ab3
movq 0x38(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x1645b0b
movq %rax, %rdi
callq 0x678a0
jmp 0x1645b0d
leaq 0x100(%rsp), %rax
movq %rax, 0x198(%rsp)
movq 0x198(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x30(%rsp)
leaq 0x100(%rsp), %rax
movq %rax, 0x1a8(%rsp)
movq 0x1a8(%rsp), %rax
movq %rax, 0x300(%rsp)
movq 0x300(%rsp), %rax
movq %rax, 0x28(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x1645bec
movq 0x28(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x2fc(%rsp) # imm = 0xFFFFFFFF
movl 0x2fc(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x2f8(%rsp)
cmpl $0x1, 0x2f8(%rsp)
jne 0x1645bec
movq 0x28(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x1645bbd
movq 0x28(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x1645bbb
jmp 0x1645bea
movq 0x28(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x320(%rsp)
cmpq $0x0, 0x320(%rsp)
je 0x1645be8
movq 0x320(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x1645bea
jmp 0x1645bec
movq 0x28(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x1645c44
movq %rax, %rdi
callq 0x678a0
movq 0x30(%rsp), %rax
movq %rax, 0x148(%rsp)
movl $0x0, 0xf0(%rsp)
movl 0xf0(%rsp), %eax
addl $0xf, %eax
cmpl 0x154(%rsp), %eax
jge 0x1645d16
movq 0x148(%rsp), %rax
movq %rax, 0x1b8(%rsp)
movq 0x1b8(%rsp), %rax
vmovups (%rax), %zmm0
vmovaps %zmm0, 0x80(%rsp)
leaq 0x16f(%rsp), %rdi
leaq 0x80(%rsp), %rsi
callq 0x164d830
vmovaps %zmm0, 0x80(%rsp)
movq 0x148(%rsp), %rax
vmovaps 0x80(%rsp), %zmm0
movq %rax, 0x230(%rsp)
vmovaps %zmm0, 0x1c0(%rsp)
vmovaps 0x1c0(%rsp), %zmm0
movq 0x230(%rsp), %rax
vmovups %zmm0, (%rax)
movq 0x148(%rsp), %rax
addq $0x40, %rax
movq %rax, 0x148(%rsp)
movl 0xf0(%rsp), %eax
addl $0x10, %eax
movl %eax, 0xf0(%rsp)
jmp 0x1645c5c
jmp 0x1645d18
movl 0xf0(%rsp), %eax
addl $0x7, %eax
cmpl 0x154(%rsp), %eax
jge 0x1645dc7
movq 0x148(%rsp), %rax
movq %rax, 0x238(%rsp)
movq 0x238(%rsp), %rax
vmovups (%rax), %ymm0
vmovaps %ymm0, 0x60(%rsp)
leaq 0x16f(%rsp), %rdi
leaq 0x60(%rsp), %rsi
callq 0x164d8e0
vmovaps %ymm0, 0x60(%rsp)
movq 0x148(%rsp), %rax
vmovaps 0x60(%rsp), %ymm0
movq %rax, 0x270(%rsp)
vmovaps %ymm0, 0x240(%rsp)
vmovaps 0x240(%rsp), %ymm0
movq 0x270(%rsp), %rax
vmovups %ymm0, (%rax)
movq 0x148(%rsp), %rax
addq $0x20, %rax
movq %rax, 0x148(%rsp)
movl 0xf0(%rsp), %eax
addl $0x8, %eax
movl %eax, 0xf0(%rsp)
jmp 0x1645d18
jmp 0x1645dc9
movl 0xf0(%rsp), %eax
addl $0x3, %eax
cmpl 0x154(%rsp), %eax
jge 0x1645e7b
movq 0x148(%rsp), %rax
movq %rax, 0x278(%rsp)
movq 0x278(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm0, 0x50(%rsp)
leaq 0x16f(%rsp), %rdi
leaq 0x50(%rsp), %rsi
vzeroupper
callq 0x164d970
vmovaps %xmm0, 0x50(%rsp)
movq 0x148(%rsp), %rax
vmovaps 0x50(%rsp), %xmm0
movq %rax, 0x298(%rsp)
vmovaps %xmm0, 0x280(%rsp)
vmovaps 0x280(%rsp), %xmm0
movq 0x298(%rsp), %rax
vmovaps %xmm0, (%rax)
movq 0x148(%rsp), %rax
addq $0x10, %rax
movq %rax, 0x148(%rsp)
movl 0xf0(%rsp), %eax
addl $0x4, %eax
movl %eax, 0xf0(%rsp)
jmp 0x1645dc9
jmp 0x1645e7d
movl 0xf0(%rsp), %eax
cmpl 0x154(%rsp), %eax
jge 0x1645ed8
movq 0x148(%rsp), %rsi
leaq 0x16f(%rsp), %rdi
vzeroupper
callq 0x164d9f0
movq 0x148(%rsp), %rax
vmovss %xmm0, (%rax)
movq 0x148(%rsp), %rax
addq $0x4, %rax
movq %rax, 0x148(%rsp)
movl 0xf0(%rsp), %eax
addl $0x1, %eax
movl %eax, 0xf0(%rsp)
jmp 0x1645e7d
jmp 0x1645eda
movl 0x150(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x150(%rsp)
jmp 0x164581a
xorl %eax, %eax
movq %rbp, %rsp
popq %rbp
vzeroupper
retq
nopw (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,236 |
int ncnn::unary_op_inplace<ncnn::UnaryOp_x86_avx512_functor::unary_op_neg>(ncnn::Mat&, ncnn::Option const&)
|
static int unary_op_inplace(Mat& a, const Option& opt)
{
Op op;
int w = a.w;
int h = a.h;
int d = a.d;
int channels = a.c;
int elempack = a.elempack;
int size = w * h * d * elempack;
#pragma omp parallel for num_threads(opt.num_threads)
for (int q = 0; q < channels; q++)
{
float* ptr = a.channel(q);
int i = 0;
#if __SSE2__
#if __AVX__
#if __AVX512F__
for (; i + 15 < size; i += 16)
{
__m512 _p = _mm512_loadu_ps(ptr);
_p = op.func_pack16(_p);
_mm512_storeu_ps(ptr, _p);
ptr += 16;
}
#endif // __AVX512F__
for (; i + 7 < size; i += 8)
{
__m256 _p = _mm256_loadu_ps(ptr);
_p = op.func_pack8(_p);
_mm256_storeu_ps(ptr, _p);
ptr += 8;
}
#endif // __AVX__
for (; i + 3 < size; i += 4)
{
__m128 _p = _mm_load_ps(ptr);
_p = op.func_pack4(_p);
_mm_store_ps(ptr, _p);
ptr += 4;
}
#endif // __SSE2__
for (; i < size; i++)
{
*ptr = op.func(*ptr);
ptr++;
}
}
return 0;
}
|
pushq %rbp
movq %rsp, %rbp
andq $-0x40, %rsp
subq $0x2c0, %rsp # imm = 0x2C0
movq %rdi, 0x140(%rsp)
movq %rsi, 0x138(%rsp)
movq 0x140(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x130(%rsp)
movq 0x140(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0x12c(%rsp)
movq 0x140(%rsp), %rax
movl 0x34(%rax), %eax
movl %eax, 0x128(%rsp)
movq 0x140(%rsp), %rax
movl 0x38(%rax), %eax
movl %eax, 0x124(%rsp)
movq 0x140(%rsp), %rax
movl 0x18(%rax), %eax
movl %eax, 0x120(%rsp)
movl 0x130(%rsp), %eax
imull 0x12c(%rsp), %eax
imull 0x128(%rsp), %eax
imull 0x120(%rsp), %eax
movl %eax, 0x11c(%rsp)
movl $0x0, 0x118(%rsp)
movl 0x118(%rsp), %eax
cmpl 0x124(%rsp), %eax
jge 0x1646680
movq 0x140(%rsp), %rcx
movl 0x118(%rsp), %eax
leaq 0xc8(%rsp), %rdx
movq %rdx, 0x158(%rsp)
movq %rcx, 0x150(%rsp)
movl %eax, 0x14c(%rsp)
movq 0x150(%rsp), %rax
movq %rax, 0x48(%rsp)
movb $0x0, 0x14b(%rsp)
movl 0x2c(%rax), %r9d
movl 0x30(%rax), %r8d
movl 0x34(%rax), %edi
movq (%rax), %rsi
movq 0x40(%rax), %rcx
movslq 0x14c(%rsp), %rdx
imulq %rdx, %rcx
imulq 0x10(%rax), %rcx
addq %rcx, %rsi
movq 0x10(%rax), %rdx
movl 0x18(%rax), %ecx
movq 0x20(%rax), %rax
leaq 0xc8(%rsp), %r10
movq %r10, 0x268(%rsp)
movl %r9d, 0x264(%rsp)
movl %r8d, 0x260(%rsp)
movl %edi, 0x25c(%rsp)
movq %rsi, 0x250(%rsp)
movq %rdx, 0x248(%rsp)
movl %ecx, 0x244(%rsp)
movq %rax, 0x238(%rsp)
movq 0x268(%rsp), %rcx
movq %rcx, 0x40(%rsp)
movq 0x250(%rsp), %rax
movq %rax, (%rcx)
movq $0x0, 0x8(%rcx)
movq 0x248(%rsp), %rax
movq %rax, 0x10(%rcx)
movl 0x244(%rsp), %eax
movl %eax, 0x18(%rcx)
movq 0x238(%rsp), %rax
movq %rax, 0x20(%rcx)
movl $0x3, 0x28(%rcx)
movl 0x264(%rsp), %eax
movl %eax, 0x2c(%rcx)
movl 0x260(%rsp), %eax
movl %eax, 0x30(%rcx)
movl $0x1, 0x34(%rcx)
movl 0x25c(%rsp), %eax
movl %eax, 0x38(%rcx)
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rdx
imulq %rdx, %rax
imulq 0x10(%rcx), %rax
movq %rax, 0x278(%rsp)
movl $0x10, 0x274(%rsp)
movq 0x278(%rsp), %rax
movslq 0x274(%rsp), %rdx
addq %rdx, %rax
subq $0x1, %rax
xorl %edx, %edx
subl 0x274(%rsp), %edx
movslq %edx, %rdx
andq %rdx, %rax
xorl %edx, %edx
divq 0x10(%rcx)
movq %rax, %rdx
movq 0x48(%rsp), %rax
movq %rdx, 0x40(%rcx)
movl 0x28(%rax), %ecx
subl $0x1, %ecx
movl %ecx, 0xf0(%rsp)
cmpl $0x4, 0x28(%rax)
jne 0x164616e
movq 0x48(%rsp), %rcx
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rcx
imulq %rcx, %rax
movq %rax, 0x108(%rsp)
movb $0x1, 0x14b(%rsp)
testb $0x1, 0x14b(%rsp)
jne 0x164629d
leaq 0xc8(%rsp), %rax
movq %rax, 0x168(%rsp)
movq 0x168(%rsp), %rax
movq %rax, 0x298(%rsp)
movq 0x298(%rsp), %rax
movq %rax, 0x38(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x1646243
movq 0x38(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x294(%rsp) # imm = 0xFFFFFFFF
movl 0x294(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x290(%rsp)
cmpl $0x1, 0x290(%rsp)
jne 0x1646243
movq 0x38(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x1646214
movq 0x38(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x1646212
jmp 0x1646241
movq 0x38(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x2a0(%rsp)
cmpq $0x0, 0x2a0(%rsp)
je 0x164623f
movq 0x2a0(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x1646241
jmp 0x1646243
movq 0x38(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x164629b
movq %rax, %rdi
callq 0x678a0
jmp 0x164629d
leaq 0xc8(%rsp), %rax
movq %rax, 0x160(%rsp)
movq 0x160(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x28(%rsp)
leaq 0xc8(%rsp), %rax
movq %rax, 0x170(%rsp)
movq 0x170(%rsp), %rax
movq %rax, 0x288(%rsp)
movq 0x288(%rsp), %rax
movq %rax, 0x30(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x164637c
movq 0x30(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x284(%rsp) # imm = 0xFFFFFFFF
movl 0x284(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x280(%rsp)
cmpl $0x1, 0x280(%rsp)
jne 0x164637c
movq 0x30(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x164634d
movq 0x30(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x164634b
jmp 0x164637a
movq 0x30(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x2a8(%rsp)
cmpq $0x0, 0x2a8(%rsp)
je 0x1646378
movq 0x2a8(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x164637a
jmp 0x164637c
movq 0x30(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x16463d4
movq %rax, %rdi
callq 0x678a0
movq 0x28(%rsp), %rax
movq %rax, 0x110(%rsp)
movl $0x0, 0xc4(%rsp)
movl 0xc4(%rsp), %eax
addl $0xf, %eax
cmpl 0x11c(%rsp), %eax
jge 0x16464a6
movq 0x110(%rsp), %rax
movq %rax, 0x178(%rsp)
movq 0x178(%rsp), %rax
vmovups (%rax), %zmm0
vmovaps %zmm0, 0x80(%rsp)
leaq 0x137(%rsp), %rdi
leaq 0x80(%rsp), %rsi
callq 0x164da20
vmovaps %zmm0, 0x80(%rsp)
movq 0x110(%rsp), %rax
vmovaps 0x80(%rsp), %zmm0
movq %rax, 0x1d0(%rsp)
vmovaps %zmm0, 0x180(%rsp)
vmovaps 0x180(%rsp), %zmm0
movq 0x1d0(%rsp), %rax
vmovups %zmm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x40, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x10, %eax
movl %eax, 0xc4(%rsp)
jmp 0x16463ec
jmp 0x16464a8
movl 0xc4(%rsp), %eax
addl $0x7, %eax
cmpl 0x11c(%rsp), %eax
jge 0x1646557
movq 0x110(%rsp), %rax
movq %rax, 0x1d8(%rsp)
movq 0x1d8(%rsp), %rax
vmovups (%rax), %ymm0
vmovaps %ymm0, 0x60(%rsp)
leaq 0x137(%rsp), %rdi
leaq 0x60(%rsp), %rsi
callq 0x164da80
vmovaps %ymm0, 0x60(%rsp)
movq 0x110(%rsp), %rax
vmovaps 0x60(%rsp), %ymm0
movq %rax, 0x210(%rsp)
vmovaps %ymm0, 0x1e0(%rsp)
vmovaps 0x1e0(%rsp), %ymm0
movq 0x210(%rsp), %rax
vmovups %ymm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x20, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x8, %eax
movl %eax, 0xc4(%rsp)
jmp 0x16464a8
jmp 0x1646559
movl 0xc4(%rsp), %eax
addl $0x3, %eax
cmpl 0x11c(%rsp), %eax
jge 0x164660b
movq 0x110(%rsp), %rax
movq %rax, 0x218(%rsp)
movq 0x218(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm0, 0x50(%rsp)
leaq 0x137(%rsp), %rdi
leaq 0x50(%rsp), %rsi
vzeroupper
callq 0x164dad0
vmovaps %xmm0, 0x50(%rsp)
movq 0x110(%rsp), %rax
vmovaps 0x50(%rsp), %xmm0
movq %rax, 0x230(%rsp)
vmovaps %xmm0, 0x220(%rsp)
vmovaps 0x220(%rsp), %xmm0
movq 0x230(%rsp), %rax
vmovaps %xmm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x10, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x4, %eax
movl %eax, 0xc4(%rsp)
jmp 0x1646559
jmp 0x164660d
movl 0xc4(%rsp), %eax
cmpl 0x11c(%rsp), %eax
jge 0x1646668
movq 0x110(%rsp), %rsi
leaq 0x137(%rsp), %rdi
vzeroupper
callq 0x164db10
movq 0x110(%rsp), %rax
vmovss %xmm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x4, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x1, %eax
movl %eax, 0xc4(%rsp)
jmp 0x164660d
jmp 0x164666a
movl 0x118(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x118(%rsp)
jmp 0x1645faa
xorl %eax, %eax
movq %rbp, %rsp
popq %rbp
vzeroupper
retq
nopw (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,237 |
int ncnn::unary_op_inplace<ncnn::UnaryOp_x86_avx512_functor::unary_op_floor>(ncnn::Mat&, ncnn::Option const&)
|
static int unary_op_inplace(Mat& a, const Option& opt)
{
Op op;
int w = a.w;
int h = a.h;
int d = a.d;
int channels = a.c;
int elempack = a.elempack;
int size = w * h * d * elempack;
#pragma omp parallel for num_threads(opt.num_threads)
for (int q = 0; q < channels; q++)
{
float* ptr = a.channel(q);
int i = 0;
#if __SSE2__
#if __AVX__
#if __AVX512F__
for (; i + 15 < size; i += 16)
{
__m512 _p = _mm512_loadu_ps(ptr);
_p = op.func_pack16(_p);
_mm512_storeu_ps(ptr, _p);
ptr += 16;
}
#endif // __AVX512F__
for (; i + 7 < size; i += 8)
{
__m256 _p = _mm256_loadu_ps(ptr);
_p = op.func_pack8(_p);
_mm256_storeu_ps(ptr, _p);
ptr += 8;
}
#endif // __AVX__
for (; i + 3 < size; i += 4)
{
__m128 _p = _mm_load_ps(ptr);
_p = op.func_pack4(_p);
_mm_store_ps(ptr, _p);
ptr += 4;
}
#endif // __SSE2__
for (; i < size; i++)
{
*ptr = op.func(*ptr);
ptr++;
}
}
return 0;
}
|
pushq %rbp
movq %rsp, %rbp
andq $-0x40, %rsp
subq $0x2c0, %rsp # imm = 0x2C0
movq %rdi, 0x140(%rsp)
movq %rsi, 0x138(%rsp)
movq 0x140(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x130(%rsp)
movq 0x140(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0x12c(%rsp)
movq 0x140(%rsp), %rax
movl 0x34(%rax), %eax
movl %eax, 0x128(%rsp)
movq 0x140(%rsp), %rax
movl 0x38(%rax), %eax
movl %eax, 0x124(%rsp)
movq 0x140(%rsp), %rax
movl 0x18(%rax), %eax
movl %eax, 0x120(%rsp)
movl 0x130(%rsp), %eax
imull 0x12c(%rsp), %eax
imull 0x128(%rsp), %eax
imull 0x120(%rsp), %eax
movl %eax, 0x11c(%rsp)
movl $0x0, 0x118(%rsp)
movl 0x118(%rsp), %eax
cmpl 0x124(%rsp), %eax
jge 0x1646e10
movq 0x140(%rsp), %rcx
movl 0x118(%rsp), %eax
leaq 0xc8(%rsp), %rdx
movq %rdx, 0x158(%rsp)
movq %rcx, 0x150(%rsp)
movl %eax, 0x14c(%rsp)
movq 0x150(%rsp), %rax
movq %rax, 0x48(%rsp)
movb $0x0, 0x14b(%rsp)
movl 0x2c(%rax), %r9d
movl 0x30(%rax), %r8d
movl 0x34(%rax), %edi
movq (%rax), %rsi
movq 0x40(%rax), %rcx
movslq 0x14c(%rsp), %rdx
imulq %rdx, %rcx
imulq 0x10(%rax), %rcx
addq %rcx, %rsi
movq 0x10(%rax), %rdx
movl 0x18(%rax), %ecx
movq 0x20(%rax), %rax
leaq 0xc8(%rsp), %r10
movq %r10, 0x268(%rsp)
movl %r9d, 0x264(%rsp)
movl %r8d, 0x260(%rsp)
movl %edi, 0x25c(%rsp)
movq %rsi, 0x250(%rsp)
movq %rdx, 0x248(%rsp)
movl %ecx, 0x244(%rsp)
movq %rax, 0x238(%rsp)
movq 0x268(%rsp), %rcx
movq %rcx, 0x40(%rsp)
movq 0x250(%rsp), %rax
movq %rax, (%rcx)
movq $0x0, 0x8(%rcx)
movq 0x248(%rsp), %rax
movq %rax, 0x10(%rcx)
movl 0x244(%rsp), %eax
movl %eax, 0x18(%rcx)
movq 0x238(%rsp), %rax
movq %rax, 0x20(%rcx)
movl $0x3, 0x28(%rcx)
movl 0x264(%rsp), %eax
movl %eax, 0x2c(%rcx)
movl 0x260(%rsp), %eax
movl %eax, 0x30(%rcx)
movl $0x1, 0x34(%rcx)
movl 0x25c(%rsp), %eax
movl %eax, 0x38(%rcx)
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rdx
imulq %rdx, %rax
imulq 0x10(%rcx), %rax
movq %rax, 0x278(%rsp)
movl $0x10, 0x274(%rsp)
movq 0x278(%rsp), %rax
movslq 0x274(%rsp), %rdx
addq %rdx, %rax
subq $0x1, %rax
xorl %edx, %edx
subl 0x274(%rsp), %edx
movslq %edx, %rdx
andq %rdx, %rax
xorl %edx, %edx
divq 0x10(%rcx)
movq %rax, %rdx
movq 0x48(%rsp), %rax
movq %rdx, 0x40(%rcx)
movl 0x28(%rax), %ecx
subl $0x1, %ecx
movl %ecx, 0xf0(%rsp)
cmpl $0x4, 0x28(%rax)
jne 0x16468fe
movq 0x48(%rsp), %rcx
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rcx
imulq %rcx, %rax
movq %rax, 0x108(%rsp)
movb $0x1, 0x14b(%rsp)
testb $0x1, 0x14b(%rsp)
jne 0x1646a2d
leaq 0xc8(%rsp), %rax
movq %rax, 0x168(%rsp)
movq 0x168(%rsp), %rax
movq %rax, 0x298(%rsp)
movq 0x298(%rsp), %rax
movq %rax, 0x38(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x16469d3
movq 0x38(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x294(%rsp) # imm = 0xFFFFFFFF
movl 0x294(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x290(%rsp)
cmpl $0x1, 0x290(%rsp)
jne 0x16469d3
movq 0x38(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x16469a4
movq 0x38(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x16469a2
jmp 0x16469d1
movq 0x38(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x2a0(%rsp)
cmpq $0x0, 0x2a0(%rsp)
je 0x16469cf
movq 0x2a0(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x16469d1
jmp 0x16469d3
movq 0x38(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x1646a2b
movq %rax, %rdi
callq 0x678a0
jmp 0x1646a2d
leaq 0xc8(%rsp), %rax
movq %rax, 0x160(%rsp)
movq 0x160(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x28(%rsp)
leaq 0xc8(%rsp), %rax
movq %rax, 0x170(%rsp)
movq 0x170(%rsp), %rax
movq %rax, 0x288(%rsp)
movq 0x288(%rsp), %rax
movq %rax, 0x30(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x1646b0c
movq 0x30(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x284(%rsp) # imm = 0xFFFFFFFF
movl 0x284(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x280(%rsp)
cmpl $0x1, 0x280(%rsp)
jne 0x1646b0c
movq 0x30(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x1646add
movq 0x30(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x1646adb
jmp 0x1646b0a
movq 0x30(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x2a8(%rsp)
cmpq $0x0, 0x2a8(%rsp)
je 0x1646b08
movq 0x2a8(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x1646b0a
jmp 0x1646b0c
movq 0x30(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x1646b64
movq %rax, %rdi
callq 0x678a0
movq 0x28(%rsp), %rax
movq %rax, 0x110(%rsp)
movl $0x0, 0xc4(%rsp)
movl 0xc4(%rsp), %eax
addl $0xf, %eax
cmpl 0x11c(%rsp), %eax
jge 0x1646c36
movq 0x110(%rsp), %rax
movq %rax, 0x178(%rsp)
movq 0x178(%rsp), %rax
vmovups (%rax), %zmm0
vmovaps %zmm0, 0x80(%rsp)
leaq 0x137(%rsp), %rdi
leaq 0x80(%rsp), %rsi
callq 0x164db40
vmovaps %zmm0, 0x80(%rsp)
movq 0x110(%rsp), %rax
vmovaps 0x80(%rsp), %zmm0
movq %rax, 0x1d0(%rsp)
vmovaps %zmm0, 0x180(%rsp)
vmovaps 0x180(%rsp), %zmm0
movq 0x1d0(%rsp), %rax
vmovups %zmm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x40, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x10, %eax
movl %eax, 0xc4(%rsp)
jmp 0x1646b7c
jmp 0x1646c38
movl 0xc4(%rsp), %eax
addl $0x7, %eax
cmpl 0x11c(%rsp), %eax
jge 0x1646ce7
movq 0x110(%rsp), %rax
movq %rax, 0x1d8(%rsp)
movq 0x1d8(%rsp), %rax
vmovups (%rax), %ymm0
vmovaps %ymm0, 0x60(%rsp)
leaq 0x137(%rsp), %rdi
leaq 0x60(%rsp), %rsi
callq 0x164db60
vmovaps %ymm0, 0x60(%rsp)
movq 0x110(%rsp), %rax
vmovaps 0x60(%rsp), %ymm0
movq %rax, 0x210(%rsp)
vmovaps %ymm0, 0x1e0(%rsp)
vmovaps 0x1e0(%rsp), %ymm0
movq 0x210(%rsp), %rax
vmovups %ymm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x20, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x8, %eax
movl %eax, 0xc4(%rsp)
jmp 0x1646c38
jmp 0x1646ce9
movl 0xc4(%rsp), %eax
addl $0x3, %eax
cmpl 0x11c(%rsp), %eax
jge 0x1646d9b
movq 0x110(%rsp), %rax
movq %rax, 0x218(%rsp)
movq 0x218(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm0, 0x50(%rsp)
leaq 0x137(%rsp), %rdi
leaq 0x50(%rsp), %rsi
vzeroupper
callq 0x164db80
vmovaps %xmm0, 0x50(%rsp)
movq 0x110(%rsp), %rax
vmovaps 0x50(%rsp), %xmm0
movq %rax, 0x230(%rsp)
vmovaps %xmm0, 0x220(%rsp)
vmovaps 0x220(%rsp), %xmm0
movq 0x230(%rsp), %rax
vmovaps %xmm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x10, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x4, %eax
movl %eax, 0xc4(%rsp)
jmp 0x1646ce9
jmp 0x1646d9d
movl 0xc4(%rsp), %eax
cmpl 0x11c(%rsp), %eax
jge 0x1646df8
movq 0x110(%rsp), %rsi
leaq 0x137(%rsp), %rdi
vzeroupper
callq 0x164dbb0
movq 0x110(%rsp), %rax
vmovss %xmm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x4, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x1, %eax
movl %eax, 0xc4(%rsp)
jmp 0x1646d9d
jmp 0x1646dfa
movl 0x118(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x118(%rsp)
jmp 0x164673a
xorl %eax, %eax
movq %rbp, %rsp
popq %rbp
vzeroupper
retq
nopw (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,238 |
int ncnn::unary_op_inplace<ncnn::UnaryOp_x86_avx512_functor::unary_op_ceil>(ncnn::Mat&, ncnn::Option const&)
|
static int unary_op_inplace(Mat& a, const Option& opt)
{
Op op;
int w = a.w;
int h = a.h;
int d = a.d;
int channels = a.c;
int elempack = a.elempack;
int size = w * h * d * elempack;
#pragma omp parallel for num_threads(opt.num_threads)
for (int q = 0; q < channels; q++)
{
float* ptr = a.channel(q);
int i = 0;
#if __SSE2__
#if __AVX__
#if __AVX512F__
for (; i + 15 < size; i += 16)
{
__m512 _p = _mm512_loadu_ps(ptr);
_p = op.func_pack16(_p);
_mm512_storeu_ps(ptr, _p);
ptr += 16;
}
#endif // __AVX512F__
for (; i + 7 < size; i += 8)
{
__m256 _p = _mm256_loadu_ps(ptr);
_p = op.func_pack8(_p);
_mm256_storeu_ps(ptr, _p);
ptr += 8;
}
#endif // __AVX__
for (; i + 3 < size; i += 4)
{
__m128 _p = _mm_load_ps(ptr);
_p = op.func_pack4(_p);
_mm_store_ps(ptr, _p);
ptr += 4;
}
#endif // __SSE2__
for (; i < size; i++)
{
*ptr = op.func(*ptr);
ptr++;
}
}
return 0;
}
|
pushq %rbp
movq %rsp, %rbp
andq $-0x40, %rsp
subq $0x2c0, %rsp # imm = 0x2C0
movq %rdi, 0x140(%rsp)
movq %rsi, 0x138(%rsp)
movq 0x140(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x130(%rsp)
movq 0x140(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0x12c(%rsp)
movq 0x140(%rsp), %rax
movl 0x34(%rax), %eax
movl %eax, 0x128(%rsp)
movq 0x140(%rsp), %rax
movl 0x38(%rax), %eax
movl %eax, 0x124(%rsp)
movq 0x140(%rsp), %rax
movl 0x18(%rax), %eax
movl %eax, 0x120(%rsp)
movl 0x130(%rsp), %eax
imull 0x12c(%rsp), %eax
imull 0x128(%rsp), %eax
imull 0x120(%rsp), %eax
movl %eax, 0x11c(%rsp)
movl $0x0, 0x118(%rsp)
movl 0x118(%rsp), %eax
cmpl 0x124(%rsp), %eax
jge 0x16475a0
movq 0x140(%rsp), %rcx
movl 0x118(%rsp), %eax
leaq 0xc8(%rsp), %rdx
movq %rdx, 0x158(%rsp)
movq %rcx, 0x150(%rsp)
movl %eax, 0x14c(%rsp)
movq 0x150(%rsp), %rax
movq %rax, 0x48(%rsp)
movb $0x0, 0x14b(%rsp)
movl 0x2c(%rax), %r9d
movl 0x30(%rax), %r8d
movl 0x34(%rax), %edi
movq (%rax), %rsi
movq 0x40(%rax), %rcx
movslq 0x14c(%rsp), %rdx
imulq %rdx, %rcx
imulq 0x10(%rax), %rcx
addq %rcx, %rsi
movq 0x10(%rax), %rdx
movl 0x18(%rax), %ecx
movq 0x20(%rax), %rax
leaq 0xc8(%rsp), %r10
movq %r10, 0x268(%rsp)
movl %r9d, 0x264(%rsp)
movl %r8d, 0x260(%rsp)
movl %edi, 0x25c(%rsp)
movq %rsi, 0x250(%rsp)
movq %rdx, 0x248(%rsp)
movl %ecx, 0x244(%rsp)
movq %rax, 0x238(%rsp)
movq 0x268(%rsp), %rcx
movq %rcx, 0x40(%rsp)
movq 0x250(%rsp), %rax
movq %rax, (%rcx)
movq $0x0, 0x8(%rcx)
movq 0x248(%rsp), %rax
movq %rax, 0x10(%rcx)
movl 0x244(%rsp), %eax
movl %eax, 0x18(%rcx)
movq 0x238(%rsp), %rax
movq %rax, 0x20(%rcx)
movl $0x3, 0x28(%rcx)
movl 0x264(%rsp), %eax
movl %eax, 0x2c(%rcx)
movl 0x260(%rsp), %eax
movl %eax, 0x30(%rcx)
movl $0x1, 0x34(%rcx)
movl 0x25c(%rsp), %eax
movl %eax, 0x38(%rcx)
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rdx
imulq %rdx, %rax
imulq 0x10(%rcx), %rax
movq %rax, 0x278(%rsp)
movl $0x10, 0x274(%rsp)
movq 0x278(%rsp), %rax
movslq 0x274(%rsp), %rdx
addq %rdx, %rax
subq $0x1, %rax
xorl %edx, %edx
subl 0x274(%rsp), %edx
movslq %edx, %rdx
andq %rdx, %rax
xorl %edx, %edx
divq 0x10(%rcx)
movq %rax, %rdx
movq 0x48(%rsp), %rax
movq %rdx, 0x40(%rcx)
movl 0x28(%rax), %ecx
subl $0x1, %ecx
movl %ecx, 0xf0(%rsp)
cmpl $0x4, 0x28(%rax)
jne 0x164708e
movq 0x48(%rsp), %rcx
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rcx
imulq %rcx, %rax
movq %rax, 0x108(%rsp)
movb $0x1, 0x14b(%rsp)
testb $0x1, 0x14b(%rsp)
jne 0x16471bd
leaq 0xc8(%rsp), %rax
movq %rax, 0x168(%rsp)
movq 0x168(%rsp), %rax
movq %rax, 0x298(%rsp)
movq 0x298(%rsp), %rax
movq %rax, 0x38(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x1647163
movq 0x38(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x294(%rsp) # imm = 0xFFFFFFFF
movl 0x294(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x290(%rsp)
cmpl $0x1, 0x290(%rsp)
jne 0x1647163
movq 0x38(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x1647134
movq 0x38(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x1647132
jmp 0x1647161
movq 0x38(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x2a0(%rsp)
cmpq $0x0, 0x2a0(%rsp)
je 0x164715f
movq 0x2a0(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x1647161
jmp 0x1647163
movq 0x38(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x16471bb
movq %rax, %rdi
callq 0x678a0
jmp 0x16471bd
leaq 0xc8(%rsp), %rax
movq %rax, 0x160(%rsp)
movq 0x160(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x28(%rsp)
leaq 0xc8(%rsp), %rax
movq %rax, 0x170(%rsp)
movq 0x170(%rsp), %rax
movq %rax, 0x288(%rsp)
movq 0x288(%rsp), %rax
movq %rax, 0x30(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x164729c
movq 0x30(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x284(%rsp) # imm = 0xFFFFFFFF
movl 0x284(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x280(%rsp)
cmpl $0x1, 0x280(%rsp)
jne 0x164729c
movq 0x30(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x164726d
movq 0x30(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x164726b
jmp 0x164729a
movq 0x30(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x2a8(%rsp)
cmpq $0x0, 0x2a8(%rsp)
je 0x1647298
movq 0x2a8(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x164729a
jmp 0x164729c
movq 0x30(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x16472f4
movq %rax, %rdi
callq 0x678a0
movq 0x28(%rsp), %rax
movq %rax, 0x110(%rsp)
movl $0x0, 0xc4(%rsp)
movl 0xc4(%rsp), %eax
addl $0xf, %eax
cmpl 0x11c(%rsp), %eax
jge 0x16473c6
movq 0x110(%rsp), %rax
movq %rax, 0x178(%rsp)
movq 0x178(%rsp), %rax
vmovups (%rax), %zmm0
vmovaps %zmm0, 0x80(%rsp)
leaq 0x137(%rsp), %rdi
leaq 0x80(%rsp), %rsi
callq 0x164dbe0
vmovaps %zmm0, 0x80(%rsp)
movq 0x110(%rsp), %rax
vmovaps 0x80(%rsp), %zmm0
movq %rax, 0x1d0(%rsp)
vmovaps %zmm0, 0x180(%rsp)
vmovaps 0x180(%rsp), %zmm0
movq 0x1d0(%rsp), %rax
vmovups %zmm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x40, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x10, %eax
movl %eax, 0xc4(%rsp)
jmp 0x164730c
jmp 0x16473c8
movl 0xc4(%rsp), %eax
addl $0x7, %eax
cmpl 0x11c(%rsp), %eax
jge 0x1647477
movq 0x110(%rsp), %rax
movq %rax, 0x1d8(%rsp)
movq 0x1d8(%rsp), %rax
vmovups (%rax), %ymm0
vmovaps %ymm0, 0x60(%rsp)
leaq 0x137(%rsp), %rdi
leaq 0x60(%rsp), %rsi
callq 0x164dc00
vmovaps %ymm0, 0x60(%rsp)
movq 0x110(%rsp), %rax
vmovaps 0x60(%rsp), %ymm0
movq %rax, 0x210(%rsp)
vmovaps %ymm0, 0x1e0(%rsp)
vmovaps 0x1e0(%rsp), %ymm0
movq 0x210(%rsp), %rax
vmovups %ymm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x20, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x8, %eax
movl %eax, 0xc4(%rsp)
jmp 0x16473c8
jmp 0x1647479
movl 0xc4(%rsp), %eax
addl $0x3, %eax
cmpl 0x11c(%rsp), %eax
jge 0x164752b
movq 0x110(%rsp), %rax
movq %rax, 0x218(%rsp)
movq 0x218(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm0, 0x50(%rsp)
leaq 0x137(%rsp), %rdi
leaq 0x50(%rsp), %rsi
vzeroupper
callq 0x164dc20
vmovaps %xmm0, 0x50(%rsp)
movq 0x110(%rsp), %rax
vmovaps 0x50(%rsp), %xmm0
movq %rax, 0x230(%rsp)
vmovaps %xmm0, 0x220(%rsp)
vmovaps 0x220(%rsp), %xmm0
movq 0x230(%rsp), %rax
vmovaps %xmm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x10, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x4, %eax
movl %eax, 0xc4(%rsp)
jmp 0x1647479
jmp 0x164752d
movl 0xc4(%rsp), %eax
cmpl 0x11c(%rsp), %eax
jge 0x1647588
movq 0x110(%rsp), %rsi
leaq 0x137(%rsp), %rdi
vzeroupper
callq 0x164dc50
movq 0x110(%rsp), %rax
vmovss %xmm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x4, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x1, %eax
movl %eax, 0xc4(%rsp)
jmp 0x164752d
jmp 0x164758a
movl 0x118(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x118(%rsp)
jmp 0x1646eca
xorl %eax, %eax
movq %rbp, %rsp
popq %rbp
vzeroupper
retq
nopw (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,239 |
int ncnn::unary_op_inplace<ncnn::UnaryOp_x86_avx512_functor::unary_op_square>(ncnn::Mat&, ncnn::Option const&)
|
static int unary_op_inplace(Mat& a, const Option& opt)
{
Op op;
int w = a.w;
int h = a.h;
int d = a.d;
int channels = a.c;
int elempack = a.elempack;
int size = w * h * d * elempack;
#pragma omp parallel for num_threads(opt.num_threads)
for (int q = 0; q < channels; q++)
{
float* ptr = a.channel(q);
int i = 0;
#if __SSE2__
#if __AVX__
#if __AVX512F__
for (; i + 15 < size; i += 16)
{
__m512 _p = _mm512_loadu_ps(ptr);
_p = op.func_pack16(_p);
_mm512_storeu_ps(ptr, _p);
ptr += 16;
}
#endif // __AVX512F__
for (; i + 7 < size; i += 8)
{
__m256 _p = _mm256_loadu_ps(ptr);
_p = op.func_pack8(_p);
_mm256_storeu_ps(ptr, _p);
ptr += 8;
}
#endif // __AVX__
for (; i + 3 < size; i += 4)
{
__m128 _p = _mm_load_ps(ptr);
_p = op.func_pack4(_p);
_mm_store_ps(ptr, _p);
ptr += 4;
}
#endif // __SSE2__
for (; i < size; i++)
{
*ptr = op.func(*ptr);
ptr++;
}
}
return 0;
}
|
pushq %rbp
movq %rsp, %rbp
andq $-0x40, %rsp
subq $0x2c0, %rsp # imm = 0x2C0
movq %rdi, 0x140(%rsp)
movq %rsi, 0x138(%rsp)
movq 0x140(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x130(%rsp)
movq 0x140(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0x12c(%rsp)
movq 0x140(%rsp), %rax
movl 0x34(%rax), %eax
movl %eax, 0x128(%rsp)
movq 0x140(%rsp), %rax
movl 0x38(%rax), %eax
movl %eax, 0x124(%rsp)
movq 0x140(%rsp), %rax
movl 0x18(%rax), %eax
movl %eax, 0x120(%rsp)
movl 0x130(%rsp), %eax
imull 0x12c(%rsp), %eax
imull 0x128(%rsp), %eax
imull 0x120(%rsp), %eax
movl %eax, 0x11c(%rsp)
movl $0x0, 0x118(%rsp)
movl 0x118(%rsp), %eax
cmpl 0x124(%rsp), %eax
jge 0x1647d30
movq 0x140(%rsp), %rcx
movl 0x118(%rsp), %eax
leaq 0xc8(%rsp), %rdx
movq %rdx, 0x158(%rsp)
movq %rcx, 0x150(%rsp)
movl %eax, 0x14c(%rsp)
movq 0x150(%rsp), %rax
movq %rax, 0x48(%rsp)
movb $0x0, 0x14b(%rsp)
movl 0x2c(%rax), %r9d
movl 0x30(%rax), %r8d
movl 0x34(%rax), %edi
movq (%rax), %rsi
movq 0x40(%rax), %rcx
movslq 0x14c(%rsp), %rdx
imulq %rdx, %rcx
imulq 0x10(%rax), %rcx
addq %rcx, %rsi
movq 0x10(%rax), %rdx
movl 0x18(%rax), %ecx
movq 0x20(%rax), %rax
leaq 0xc8(%rsp), %r10
movq %r10, 0x268(%rsp)
movl %r9d, 0x264(%rsp)
movl %r8d, 0x260(%rsp)
movl %edi, 0x25c(%rsp)
movq %rsi, 0x250(%rsp)
movq %rdx, 0x248(%rsp)
movl %ecx, 0x244(%rsp)
movq %rax, 0x238(%rsp)
movq 0x268(%rsp), %rcx
movq %rcx, 0x40(%rsp)
movq 0x250(%rsp), %rax
movq %rax, (%rcx)
movq $0x0, 0x8(%rcx)
movq 0x248(%rsp), %rax
movq %rax, 0x10(%rcx)
movl 0x244(%rsp), %eax
movl %eax, 0x18(%rcx)
movq 0x238(%rsp), %rax
movq %rax, 0x20(%rcx)
movl $0x3, 0x28(%rcx)
movl 0x264(%rsp), %eax
movl %eax, 0x2c(%rcx)
movl 0x260(%rsp), %eax
movl %eax, 0x30(%rcx)
movl $0x1, 0x34(%rcx)
movl 0x25c(%rsp), %eax
movl %eax, 0x38(%rcx)
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rdx
imulq %rdx, %rax
imulq 0x10(%rcx), %rax
movq %rax, 0x278(%rsp)
movl $0x10, 0x274(%rsp)
movq 0x278(%rsp), %rax
movslq 0x274(%rsp), %rdx
addq %rdx, %rax
subq $0x1, %rax
xorl %edx, %edx
subl 0x274(%rsp), %edx
movslq %edx, %rdx
andq %rdx, %rax
xorl %edx, %edx
divq 0x10(%rcx)
movq %rax, %rdx
movq 0x48(%rsp), %rax
movq %rdx, 0x40(%rcx)
movl 0x28(%rax), %ecx
subl $0x1, %ecx
movl %ecx, 0xf0(%rsp)
cmpl $0x4, 0x28(%rax)
jne 0x164781e
movq 0x48(%rsp), %rcx
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rcx
imulq %rcx, %rax
movq %rax, 0x108(%rsp)
movb $0x1, 0x14b(%rsp)
testb $0x1, 0x14b(%rsp)
jne 0x164794d
leaq 0xc8(%rsp), %rax
movq %rax, 0x168(%rsp)
movq 0x168(%rsp), %rax
movq %rax, 0x298(%rsp)
movq 0x298(%rsp), %rax
movq %rax, 0x38(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x16478f3
movq 0x38(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x294(%rsp) # imm = 0xFFFFFFFF
movl 0x294(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x290(%rsp)
cmpl $0x1, 0x290(%rsp)
jne 0x16478f3
movq 0x38(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x16478c4
movq 0x38(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x16478c2
jmp 0x16478f1
movq 0x38(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x2a0(%rsp)
cmpq $0x0, 0x2a0(%rsp)
je 0x16478ef
movq 0x2a0(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x16478f1
jmp 0x16478f3
movq 0x38(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x164794b
movq %rax, %rdi
callq 0x678a0
jmp 0x164794d
leaq 0xc8(%rsp), %rax
movq %rax, 0x160(%rsp)
movq 0x160(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x28(%rsp)
leaq 0xc8(%rsp), %rax
movq %rax, 0x170(%rsp)
movq 0x170(%rsp), %rax
movq %rax, 0x288(%rsp)
movq 0x288(%rsp), %rax
movq %rax, 0x30(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x1647a2c
movq 0x30(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x284(%rsp) # imm = 0xFFFFFFFF
movl 0x284(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x280(%rsp)
cmpl $0x1, 0x280(%rsp)
jne 0x1647a2c
movq 0x30(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x16479fd
movq 0x30(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x16479fb
jmp 0x1647a2a
movq 0x30(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x2a8(%rsp)
cmpq $0x0, 0x2a8(%rsp)
je 0x1647a28
movq 0x2a8(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x1647a2a
jmp 0x1647a2c
movq 0x30(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x1647a84
movq %rax, %rdi
callq 0x678a0
movq 0x28(%rsp), %rax
movq %rax, 0x110(%rsp)
movl $0x0, 0xc4(%rsp)
movl 0xc4(%rsp), %eax
addl $0xf, %eax
cmpl 0x11c(%rsp), %eax
jge 0x1647b56
movq 0x110(%rsp), %rax
movq %rax, 0x178(%rsp)
movq 0x178(%rsp), %rax
vmovups (%rax), %zmm0
vmovaps %zmm0, 0x80(%rsp)
leaq 0x137(%rsp), %rdi
leaq 0x80(%rsp), %rsi
callq 0x164dc80
vmovaps %zmm0, 0x80(%rsp)
movq 0x110(%rsp), %rax
vmovaps 0x80(%rsp), %zmm0
movq %rax, 0x1d0(%rsp)
vmovaps %zmm0, 0x180(%rsp)
vmovaps 0x180(%rsp), %zmm0
movq 0x1d0(%rsp), %rax
vmovups %zmm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x40, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x10, %eax
movl %eax, 0xc4(%rsp)
jmp 0x1647a9c
jmp 0x1647b58
movl 0xc4(%rsp), %eax
addl $0x7, %eax
cmpl 0x11c(%rsp), %eax
jge 0x1647c07
movq 0x110(%rsp), %rax
movq %rax, 0x1d8(%rsp)
movq 0x1d8(%rsp), %rax
vmovups (%rax), %ymm0
vmovaps %ymm0, 0x60(%rsp)
leaq 0x137(%rsp), %rdi
leaq 0x60(%rsp), %rsi
callq 0x164dce0
vmovaps %ymm0, 0x60(%rsp)
movq 0x110(%rsp), %rax
vmovaps 0x60(%rsp), %ymm0
movq %rax, 0x210(%rsp)
vmovaps %ymm0, 0x1e0(%rsp)
vmovaps 0x1e0(%rsp), %ymm0
movq 0x210(%rsp), %rax
vmovups %ymm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x20, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x8, %eax
movl %eax, 0xc4(%rsp)
jmp 0x1647b58
jmp 0x1647c09
movl 0xc4(%rsp), %eax
addl $0x3, %eax
cmpl 0x11c(%rsp), %eax
jge 0x1647cbb
movq 0x110(%rsp), %rax
movq %rax, 0x218(%rsp)
movq 0x218(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm0, 0x50(%rsp)
leaq 0x137(%rsp), %rdi
leaq 0x50(%rsp), %rsi
vzeroupper
callq 0x164dd30
vmovaps %xmm0, 0x50(%rsp)
movq 0x110(%rsp), %rax
vmovaps 0x50(%rsp), %xmm0
movq %rax, 0x230(%rsp)
vmovaps %xmm0, 0x220(%rsp)
vmovaps 0x220(%rsp), %xmm0
movq 0x230(%rsp), %rax
vmovaps %xmm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x10, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x4, %eax
movl %eax, 0xc4(%rsp)
jmp 0x1647c09
jmp 0x1647cbd
movl 0xc4(%rsp), %eax
cmpl 0x11c(%rsp), %eax
jge 0x1647d18
movq 0x110(%rsp), %rsi
leaq 0x137(%rsp), %rdi
vzeroupper
callq 0x164dd70
movq 0x110(%rsp), %rax
vmovss %xmm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x4, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x1, %eax
movl %eax, 0xc4(%rsp)
jmp 0x1647cbd
jmp 0x1647d1a
movl 0x118(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x118(%rsp)
jmp 0x164765a
xorl %eax, %eax
movq %rbp, %rsp
popq %rbp
vzeroupper
retq
nopw (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,240 |
int ncnn::unary_op_inplace<ncnn::UnaryOp_x86_avx512_functor::unary_op_sqrt>(ncnn::Mat&, ncnn::Option const&)
|
static int unary_op_inplace(Mat& a, const Option& opt)
{
Op op;
int w = a.w;
int h = a.h;
int d = a.d;
int channels = a.c;
int elempack = a.elempack;
int size = w * h * d * elempack;
#pragma omp parallel for num_threads(opt.num_threads)
for (int q = 0; q < channels; q++)
{
float* ptr = a.channel(q);
int i = 0;
#if __SSE2__
#if __AVX__
#if __AVX512F__
for (; i + 15 < size; i += 16)
{
__m512 _p = _mm512_loadu_ps(ptr);
_p = op.func_pack16(_p);
_mm512_storeu_ps(ptr, _p);
ptr += 16;
}
#endif // __AVX512F__
for (; i + 7 < size; i += 8)
{
__m256 _p = _mm256_loadu_ps(ptr);
_p = op.func_pack8(_p);
_mm256_storeu_ps(ptr, _p);
ptr += 8;
}
#endif // __AVX__
for (; i + 3 < size; i += 4)
{
__m128 _p = _mm_load_ps(ptr);
_p = op.func_pack4(_p);
_mm_store_ps(ptr, _p);
ptr += 4;
}
#endif // __SSE2__
for (; i < size; i++)
{
*ptr = op.func(*ptr);
ptr++;
}
}
return 0;
}
|
pushq %rbp
movq %rsp, %rbp
andq $-0x40, %rsp
subq $0x2c0, %rsp # imm = 0x2C0
movq %rdi, 0x140(%rsp)
movq %rsi, 0x138(%rsp)
movq 0x140(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x130(%rsp)
movq 0x140(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0x12c(%rsp)
movq 0x140(%rsp), %rax
movl 0x34(%rax), %eax
movl %eax, 0x128(%rsp)
movq 0x140(%rsp), %rax
movl 0x38(%rax), %eax
movl %eax, 0x124(%rsp)
movq 0x140(%rsp), %rax
movl 0x18(%rax), %eax
movl %eax, 0x120(%rsp)
movl 0x130(%rsp), %eax
imull 0x12c(%rsp), %eax
imull 0x128(%rsp), %eax
imull 0x120(%rsp), %eax
movl %eax, 0x11c(%rsp)
movl $0x0, 0x118(%rsp)
movl 0x118(%rsp), %eax
cmpl 0x124(%rsp), %eax
jge 0x16484c0
movq 0x140(%rsp), %rcx
movl 0x118(%rsp), %eax
leaq 0xc8(%rsp), %rdx
movq %rdx, 0x158(%rsp)
movq %rcx, 0x150(%rsp)
movl %eax, 0x14c(%rsp)
movq 0x150(%rsp), %rax
movq %rax, 0x48(%rsp)
movb $0x0, 0x14b(%rsp)
movl 0x2c(%rax), %r9d
movl 0x30(%rax), %r8d
movl 0x34(%rax), %edi
movq (%rax), %rsi
movq 0x40(%rax), %rcx
movslq 0x14c(%rsp), %rdx
imulq %rdx, %rcx
imulq 0x10(%rax), %rcx
addq %rcx, %rsi
movq 0x10(%rax), %rdx
movl 0x18(%rax), %ecx
movq 0x20(%rax), %rax
leaq 0xc8(%rsp), %r10
movq %r10, 0x268(%rsp)
movl %r9d, 0x264(%rsp)
movl %r8d, 0x260(%rsp)
movl %edi, 0x25c(%rsp)
movq %rsi, 0x250(%rsp)
movq %rdx, 0x248(%rsp)
movl %ecx, 0x244(%rsp)
movq %rax, 0x238(%rsp)
movq 0x268(%rsp), %rcx
movq %rcx, 0x40(%rsp)
movq 0x250(%rsp), %rax
movq %rax, (%rcx)
movq $0x0, 0x8(%rcx)
movq 0x248(%rsp), %rax
movq %rax, 0x10(%rcx)
movl 0x244(%rsp), %eax
movl %eax, 0x18(%rcx)
movq 0x238(%rsp), %rax
movq %rax, 0x20(%rcx)
movl $0x3, 0x28(%rcx)
movl 0x264(%rsp), %eax
movl %eax, 0x2c(%rcx)
movl 0x260(%rsp), %eax
movl %eax, 0x30(%rcx)
movl $0x1, 0x34(%rcx)
movl 0x25c(%rsp), %eax
movl %eax, 0x38(%rcx)
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rdx
imulq %rdx, %rax
imulq 0x10(%rcx), %rax
movq %rax, 0x278(%rsp)
movl $0x10, 0x274(%rsp)
movq 0x278(%rsp), %rax
movslq 0x274(%rsp), %rdx
addq %rdx, %rax
subq $0x1, %rax
xorl %edx, %edx
subl 0x274(%rsp), %edx
movslq %edx, %rdx
andq %rdx, %rax
xorl %edx, %edx
divq 0x10(%rcx)
movq %rax, %rdx
movq 0x48(%rsp), %rax
movq %rdx, 0x40(%rcx)
movl 0x28(%rax), %ecx
subl $0x1, %ecx
movl %ecx, 0xf0(%rsp)
cmpl $0x4, 0x28(%rax)
jne 0x1647fae
movq 0x48(%rsp), %rcx
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rcx
imulq %rcx, %rax
movq %rax, 0x108(%rsp)
movb $0x1, 0x14b(%rsp)
testb $0x1, 0x14b(%rsp)
jne 0x16480dd
leaq 0xc8(%rsp), %rax
movq %rax, 0x168(%rsp)
movq 0x168(%rsp), %rax
movq %rax, 0x298(%rsp)
movq 0x298(%rsp), %rax
movq %rax, 0x38(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x1648083
movq 0x38(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x294(%rsp) # imm = 0xFFFFFFFF
movl 0x294(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x290(%rsp)
cmpl $0x1, 0x290(%rsp)
jne 0x1648083
movq 0x38(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x1648054
movq 0x38(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x1648052
jmp 0x1648081
movq 0x38(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x2a0(%rsp)
cmpq $0x0, 0x2a0(%rsp)
je 0x164807f
movq 0x2a0(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x1648081
jmp 0x1648083
movq 0x38(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x16480db
movq %rax, %rdi
callq 0x678a0
jmp 0x16480dd
leaq 0xc8(%rsp), %rax
movq %rax, 0x160(%rsp)
movq 0x160(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x28(%rsp)
leaq 0xc8(%rsp), %rax
movq %rax, 0x170(%rsp)
movq 0x170(%rsp), %rax
movq %rax, 0x288(%rsp)
movq 0x288(%rsp), %rax
movq %rax, 0x30(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x16481bc
movq 0x30(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x284(%rsp) # imm = 0xFFFFFFFF
movl 0x284(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x280(%rsp)
cmpl $0x1, 0x280(%rsp)
jne 0x16481bc
movq 0x30(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x164818d
movq 0x30(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x164818b
jmp 0x16481ba
movq 0x30(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x2a8(%rsp)
cmpq $0x0, 0x2a8(%rsp)
je 0x16481b8
movq 0x2a8(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x16481ba
jmp 0x16481bc
movq 0x30(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x1648214
movq %rax, %rdi
callq 0x678a0
movq 0x28(%rsp), %rax
movq %rax, 0x110(%rsp)
movl $0x0, 0xc4(%rsp)
movl 0xc4(%rsp), %eax
addl $0xf, %eax
cmpl 0x11c(%rsp), %eax
jge 0x16482e6
movq 0x110(%rsp), %rax
movq %rax, 0x178(%rsp)
movq 0x178(%rsp), %rax
vmovups (%rax), %zmm0
vmovaps %zmm0, 0x80(%rsp)
leaq 0x137(%rsp), %rdi
leaq 0x80(%rsp), %rsi
callq 0x164dd90
vmovaps %zmm0, 0x80(%rsp)
movq 0x110(%rsp), %rax
vmovaps 0x80(%rsp), %zmm0
movq %rax, 0x1d0(%rsp)
vmovaps %zmm0, 0x180(%rsp)
vmovaps 0x180(%rsp), %zmm0
movq 0x1d0(%rsp), %rax
vmovups %zmm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x40, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x10, %eax
movl %eax, 0xc4(%rsp)
jmp 0x164822c
jmp 0x16482e8
movl 0xc4(%rsp), %eax
addl $0x7, %eax
cmpl 0x11c(%rsp), %eax
jge 0x1648397
movq 0x110(%rsp), %rax
movq %rax, 0x1d8(%rsp)
movq 0x1d8(%rsp), %rax
vmovups (%rax), %ymm0
vmovaps %ymm0, 0x60(%rsp)
leaq 0x137(%rsp), %rdi
leaq 0x60(%rsp), %rsi
callq 0x164de20
vmovaps %ymm0, 0x60(%rsp)
movq 0x110(%rsp), %rax
vmovaps 0x60(%rsp), %ymm0
movq %rax, 0x210(%rsp)
vmovaps %ymm0, 0x1e0(%rsp)
vmovaps 0x1e0(%rsp), %ymm0
movq 0x210(%rsp), %rax
vmovups %ymm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x20, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x8, %eax
movl %eax, 0xc4(%rsp)
jmp 0x16482e8
jmp 0x1648399
movl 0xc4(%rsp), %eax
addl $0x3, %eax
cmpl 0x11c(%rsp), %eax
jge 0x164844b
movq 0x110(%rsp), %rax
movq %rax, 0x218(%rsp)
movq 0x218(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm0, 0x50(%rsp)
leaq 0x137(%rsp), %rdi
leaq 0x50(%rsp), %rsi
vzeroupper
callq 0x164dea0
vmovaps %xmm0, 0x50(%rsp)
movq 0x110(%rsp), %rax
vmovaps 0x50(%rsp), %xmm0
movq %rax, 0x230(%rsp)
vmovaps %xmm0, 0x220(%rsp)
vmovaps 0x220(%rsp), %xmm0
movq 0x230(%rsp), %rax
vmovaps %xmm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x10, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x4, %eax
movl %eax, 0xc4(%rsp)
jmp 0x1648399
jmp 0x164844d
movl 0xc4(%rsp), %eax
cmpl 0x11c(%rsp), %eax
jge 0x16484a8
movq 0x110(%rsp), %rsi
leaq 0x137(%rsp), %rdi
vzeroupper
callq 0x164df10
movq 0x110(%rsp), %rax
vmovss %xmm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x4, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x1, %eax
movl %eax, 0xc4(%rsp)
jmp 0x164844d
jmp 0x16484aa
movl 0x118(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x118(%rsp)
jmp 0x1647dea
xorl %eax, %eax
movq %rbp, %rsp
popq %rbp
vzeroupper
retq
nopw (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,241 |
int ncnn::unary_op_inplace<ncnn::UnaryOp_x86_avx512_functor::unary_op_rsqrt>(ncnn::Mat&, ncnn::Option const&)
|
static int unary_op_inplace(Mat& a, const Option& opt)
{
Op op;
int w = a.w;
int h = a.h;
int d = a.d;
int channels = a.c;
int elempack = a.elempack;
int size = w * h * d * elempack;
#pragma omp parallel for num_threads(opt.num_threads)
for (int q = 0; q < channels; q++)
{
float* ptr = a.channel(q);
int i = 0;
#if __SSE2__
#if __AVX__
#if __AVX512F__
for (; i + 15 < size; i += 16)
{
__m512 _p = _mm512_loadu_ps(ptr);
_p = op.func_pack16(_p);
_mm512_storeu_ps(ptr, _p);
ptr += 16;
}
#endif // __AVX512F__
for (; i + 7 < size; i += 8)
{
__m256 _p = _mm256_loadu_ps(ptr);
_p = op.func_pack8(_p);
_mm256_storeu_ps(ptr, _p);
ptr += 8;
}
#endif // __AVX__
for (; i + 3 < size; i += 4)
{
__m128 _p = _mm_load_ps(ptr);
_p = op.func_pack4(_p);
_mm_store_ps(ptr, _p);
ptr += 4;
}
#endif // __SSE2__
for (; i < size; i++)
{
*ptr = op.func(*ptr);
ptr++;
}
}
return 0;
}
|
pushq %rbp
movq %rsp, %rbp
andq $-0x40, %rsp
subq $0x2c0, %rsp # imm = 0x2C0
movq %rdi, 0x140(%rsp)
movq %rsi, 0x138(%rsp)
movq 0x140(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x130(%rsp)
movq 0x140(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0x12c(%rsp)
movq 0x140(%rsp), %rax
movl 0x34(%rax), %eax
movl %eax, 0x128(%rsp)
movq 0x140(%rsp), %rax
movl 0x38(%rax), %eax
movl %eax, 0x124(%rsp)
movq 0x140(%rsp), %rax
movl 0x18(%rax), %eax
movl %eax, 0x120(%rsp)
movl 0x130(%rsp), %eax
imull 0x12c(%rsp), %eax
imull 0x128(%rsp), %eax
imull 0x120(%rsp), %eax
movl %eax, 0x11c(%rsp)
movl $0x0, 0x118(%rsp)
movl 0x118(%rsp), %eax
cmpl 0x124(%rsp), %eax
jge 0x1648c50
movq 0x140(%rsp), %rcx
movl 0x118(%rsp), %eax
leaq 0xc8(%rsp), %rdx
movq %rdx, 0x158(%rsp)
movq %rcx, 0x150(%rsp)
movl %eax, 0x14c(%rsp)
movq 0x150(%rsp), %rax
movq %rax, 0x48(%rsp)
movb $0x0, 0x14b(%rsp)
movl 0x2c(%rax), %r9d
movl 0x30(%rax), %r8d
movl 0x34(%rax), %edi
movq (%rax), %rsi
movq 0x40(%rax), %rcx
movslq 0x14c(%rsp), %rdx
imulq %rdx, %rcx
imulq 0x10(%rax), %rcx
addq %rcx, %rsi
movq 0x10(%rax), %rdx
movl 0x18(%rax), %ecx
movq 0x20(%rax), %rax
leaq 0xc8(%rsp), %r10
movq %r10, 0x268(%rsp)
movl %r9d, 0x264(%rsp)
movl %r8d, 0x260(%rsp)
movl %edi, 0x25c(%rsp)
movq %rsi, 0x250(%rsp)
movq %rdx, 0x248(%rsp)
movl %ecx, 0x244(%rsp)
movq %rax, 0x238(%rsp)
movq 0x268(%rsp), %rcx
movq %rcx, 0x40(%rsp)
movq 0x250(%rsp), %rax
movq %rax, (%rcx)
movq $0x0, 0x8(%rcx)
movq 0x248(%rsp), %rax
movq %rax, 0x10(%rcx)
movl 0x244(%rsp), %eax
movl %eax, 0x18(%rcx)
movq 0x238(%rsp), %rax
movq %rax, 0x20(%rcx)
movl $0x3, 0x28(%rcx)
movl 0x264(%rsp), %eax
movl %eax, 0x2c(%rcx)
movl 0x260(%rsp), %eax
movl %eax, 0x30(%rcx)
movl $0x1, 0x34(%rcx)
movl 0x25c(%rsp), %eax
movl %eax, 0x38(%rcx)
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rdx
imulq %rdx, %rax
imulq 0x10(%rcx), %rax
movq %rax, 0x278(%rsp)
movl $0x10, 0x274(%rsp)
movq 0x278(%rsp), %rax
movslq 0x274(%rsp), %rdx
addq %rdx, %rax
subq $0x1, %rax
xorl %edx, %edx
subl 0x274(%rsp), %edx
movslq %edx, %rdx
andq %rdx, %rax
xorl %edx, %edx
divq 0x10(%rcx)
movq %rax, %rdx
movq 0x48(%rsp), %rax
movq %rdx, 0x40(%rcx)
movl 0x28(%rax), %ecx
subl $0x1, %ecx
movl %ecx, 0xf0(%rsp)
cmpl $0x4, 0x28(%rax)
jne 0x164873e
movq 0x48(%rsp), %rcx
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rcx
imulq %rcx, %rax
movq %rax, 0x108(%rsp)
movb $0x1, 0x14b(%rsp)
testb $0x1, 0x14b(%rsp)
jne 0x164886d
leaq 0xc8(%rsp), %rax
movq %rax, 0x168(%rsp)
movq 0x168(%rsp), %rax
movq %rax, 0x298(%rsp)
movq 0x298(%rsp), %rax
movq %rax, 0x38(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x1648813
movq 0x38(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x294(%rsp) # imm = 0xFFFFFFFF
movl 0x294(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x290(%rsp)
cmpl $0x1, 0x290(%rsp)
jne 0x1648813
movq 0x38(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x16487e4
movq 0x38(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x16487e2
jmp 0x1648811
movq 0x38(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x2a0(%rsp)
cmpq $0x0, 0x2a0(%rsp)
je 0x164880f
movq 0x2a0(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x1648811
jmp 0x1648813
movq 0x38(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x164886b
movq %rax, %rdi
callq 0x678a0
jmp 0x164886d
leaq 0xc8(%rsp), %rax
movq %rax, 0x160(%rsp)
movq 0x160(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x28(%rsp)
leaq 0xc8(%rsp), %rax
movq %rax, 0x170(%rsp)
movq 0x170(%rsp), %rax
movq %rax, 0x288(%rsp)
movq 0x288(%rsp), %rax
movq %rax, 0x30(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x164894c
movq 0x30(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x284(%rsp) # imm = 0xFFFFFFFF
movl 0x284(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x280(%rsp)
cmpl $0x1, 0x280(%rsp)
jne 0x164894c
movq 0x30(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x164891d
movq 0x30(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x164891b
jmp 0x164894a
movq 0x30(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x2a8(%rsp)
cmpq $0x0, 0x2a8(%rsp)
je 0x1648948
movq 0x2a8(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x164894a
jmp 0x164894c
movq 0x30(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x16489a4
movq %rax, %rdi
callq 0x678a0
movq 0x28(%rsp), %rax
movq %rax, 0x110(%rsp)
movl $0x0, 0xc4(%rsp)
movl 0xc4(%rsp), %eax
addl $0xf, %eax
cmpl 0x11c(%rsp), %eax
jge 0x1648a76
movq 0x110(%rsp), %rax
movq %rax, 0x178(%rsp)
movq 0x178(%rsp), %rax
vmovups (%rax), %zmm0
vmovaps %zmm0, 0x80(%rsp)
leaq 0x137(%rsp), %rdi
leaq 0x80(%rsp), %rsi
callq 0x164df40
vmovaps %zmm0, 0x80(%rsp)
movq 0x110(%rsp), %rax
vmovaps 0x80(%rsp), %zmm0
movq %rax, 0x1d0(%rsp)
vmovaps %zmm0, 0x180(%rsp)
vmovaps 0x180(%rsp), %zmm0
movq 0x1d0(%rsp), %rax
vmovups %zmm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x40, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x10, %eax
movl %eax, 0xc4(%rsp)
jmp 0x16489bc
jmp 0x1648a78
movl 0xc4(%rsp), %eax
addl $0x7, %eax
cmpl 0x11c(%rsp), %eax
jge 0x1648b27
movq 0x110(%rsp), %rax
movq %rax, 0x1d8(%rsp)
movq 0x1d8(%rsp), %rax
vmovups (%rax), %ymm0
vmovaps %ymm0, 0x60(%rsp)
leaq 0x137(%rsp), %rdi
leaq 0x60(%rsp), %rsi
callq 0x164dfe0
vmovaps %ymm0, 0x60(%rsp)
movq 0x110(%rsp), %rax
vmovaps 0x60(%rsp), %ymm0
movq %rax, 0x210(%rsp)
vmovaps %ymm0, 0x1e0(%rsp)
vmovaps 0x1e0(%rsp), %ymm0
movq 0x210(%rsp), %rax
vmovups %ymm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x20, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x8, %eax
movl %eax, 0xc4(%rsp)
jmp 0x1648a78
jmp 0x1648b29
movl 0xc4(%rsp), %eax
addl $0x3, %eax
cmpl 0x11c(%rsp), %eax
jge 0x1648bdb
movq 0x110(%rsp), %rax
movq %rax, 0x218(%rsp)
movq 0x218(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm0, 0x50(%rsp)
leaq 0x137(%rsp), %rdi
leaq 0x50(%rsp), %rsi
vzeroupper
callq 0x164e020
vmovaps %xmm0, 0x50(%rsp)
movq 0x110(%rsp), %rax
vmovaps 0x50(%rsp), %xmm0
movq %rax, 0x230(%rsp)
vmovaps %xmm0, 0x220(%rsp)
vmovaps 0x220(%rsp), %xmm0
movq 0x230(%rsp), %rax
vmovaps %xmm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x10, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x4, %eax
movl %eax, 0xc4(%rsp)
jmp 0x1648b29
jmp 0x1648bdd
movl 0xc4(%rsp), %eax
cmpl 0x11c(%rsp), %eax
jge 0x1648c38
movq 0x110(%rsp), %rsi
leaq 0x137(%rsp), %rdi
vzeroupper
callq 0x164e050
movq 0x110(%rsp), %rax
vmovss %xmm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x4, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x1, %eax
movl %eax, 0xc4(%rsp)
jmp 0x1648bdd
jmp 0x1648c3a
movl 0x118(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x118(%rsp)
jmp 0x164857a
xorl %eax, %eax
movq %rbp, %rsp
popq %rbp
vzeroupper
retq
nopw (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,242 |
int ncnn::unary_op_inplace<ncnn::UnaryOp_x86_avx512_functor::unary_op_exp>(ncnn::Mat&, ncnn::Option const&)
|
static int unary_op_inplace(Mat& a, const Option& opt)
{
Op op;
int w = a.w;
int h = a.h;
int d = a.d;
int channels = a.c;
int elempack = a.elempack;
int size = w * h * d * elempack;
#pragma omp parallel for num_threads(opt.num_threads)
for (int q = 0; q < channels; q++)
{
float* ptr = a.channel(q);
int i = 0;
#if __SSE2__
#if __AVX__
#if __AVX512F__
for (; i + 15 < size; i += 16)
{
__m512 _p = _mm512_loadu_ps(ptr);
_p = op.func_pack16(_p);
_mm512_storeu_ps(ptr, _p);
ptr += 16;
}
#endif // __AVX512F__
for (; i + 7 < size; i += 8)
{
__m256 _p = _mm256_loadu_ps(ptr);
_p = op.func_pack8(_p);
_mm256_storeu_ps(ptr, _p);
ptr += 8;
}
#endif // __AVX__
for (; i + 3 < size; i += 4)
{
__m128 _p = _mm_load_ps(ptr);
_p = op.func_pack4(_p);
_mm_store_ps(ptr, _p);
ptr += 4;
}
#endif // __SSE2__
for (; i < size; i++)
{
*ptr = op.func(*ptr);
ptr++;
}
}
return 0;
}
|
pushq %rbp
movq %rsp, %rbp
andq $-0x40, %rsp
subq $0x2c0, %rsp # imm = 0x2C0
movq %rdi, 0x140(%rsp)
movq %rsi, 0x138(%rsp)
movq 0x140(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x130(%rsp)
movq 0x140(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0x12c(%rsp)
movq 0x140(%rsp), %rax
movl 0x34(%rax), %eax
movl %eax, 0x128(%rsp)
movq 0x140(%rsp), %rax
movl 0x38(%rax), %eax
movl %eax, 0x124(%rsp)
movq 0x140(%rsp), %rax
movl 0x18(%rax), %eax
movl %eax, 0x120(%rsp)
movl 0x130(%rsp), %eax
imull 0x12c(%rsp), %eax
imull 0x128(%rsp), %eax
imull 0x120(%rsp), %eax
movl %eax, 0x11c(%rsp)
movl $0x0, 0x118(%rsp)
movl 0x118(%rsp), %eax
cmpl 0x124(%rsp), %eax
jge 0x16493e0
movq 0x140(%rsp), %rcx
movl 0x118(%rsp), %eax
leaq 0xc8(%rsp), %rdx
movq %rdx, 0x158(%rsp)
movq %rcx, 0x150(%rsp)
movl %eax, 0x14c(%rsp)
movq 0x150(%rsp), %rax
movq %rax, 0x48(%rsp)
movb $0x0, 0x14b(%rsp)
movl 0x2c(%rax), %r9d
movl 0x30(%rax), %r8d
movl 0x34(%rax), %edi
movq (%rax), %rsi
movq 0x40(%rax), %rcx
movslq 0x14c(%rsp), %rdx
imulq %rdx, %rcx
imulq 0x10(%rax), %rcx
addq %rcx, %rsi
movq 0x10(%rax), %rdx
movl 0x18(%rax), %ecx
movq 0x20(%rax), %rax
leaq 0xc8(%rsp), %r10
movq %r10, 0x268(%rsp)
movl %r9d, 0x264(%rsp)
movl %r8d, 0x260(%rsp)
movl %edi, 0x25c(%rsp)
movq %rsi, 0x250(%rsp)
movq %rdx, 0x248(%rsp)
movl %ecx, 0x244(%rsp)
movq %rax, 0x238(%rsp)
movq 0x268(%rsp), %rcx
movq %rcx, 0x40(%rsp)
movq 0x250(%rsp), %rax
movq %rax, (%rcx)
movq $0x0, 0x8(%rcx)
movq 0x248(%rsp), %rax
movq %rax, 0x10(%rcx)
movl 0x244(%rsp), %eax
movl %eax, 0x18(%rcx)
movq 0x238(%rsp), %rax
movq %rax, 0x20(%rcx)
movl $0x3, 0x28(%rcx)
movl 0x264(%rsp), %eax
movl %eax, 0x2c(%rcx)
movl 0x260(%rsp), %eax
movl %eax, 0x30(%rcx)
movl $0x1, 0x34(%rcx)
movl 0x25c(%rsp), %eax
movl %eax, 0x38(%rcx)
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rdx
imulq %rdx, %rax
imulq 0x10(%rcx), %rax
movq %rax, 0x278(%rsp)
movl $0x10, 0x274(%rsp)
movq 0x278(%rsp), %rax
movslq 0x274(%rsp), %rdx
addq %rdx, %rax
subq $0x1, %rax
xorl %edx, %edx
subl 0x274(%rsp), %edx
movslq %edx, %rdx
andq %rdx, %rax
xorl %edx, %edx
divq 0x10(%rcx)
movq %rax, %rdx
movq 0x48(%rsp), %rax
movq %rdx, 0x40(%rcx)
movl 0x28(%rax), %ecx
subl $0x1, %ecx
movl %ecx, 0xf0(%rsp)
cmpl $0x4, 0x28(%rax)
jne 0x1648ece
movq 0x48(%rsp), %rcx
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rcx
imulq %rcx, %rax
movq %rax, 0x108(%rsp)
movb $0x1, 0x14b(%rsp)
testb $0x1, 0x14b(%rsp)
jne 0x1648ffd
leaq 0xc8(%rsp), %rax
movq %rax, 0x168(%rsp)
movq 0x168(%rsp), %rax
movq %rax, 0x298(%rsp)
movq 0x298(%rsp), %rax
movq %rax, 0x38(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x1648fa3
movq 0x38(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x294(%rsp) # imm = 0xFFFFFFFF
movl 0x294(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x290(%rsp)
cmpl $0x1, 0x290(%rsp)
jne 0x1648fa3
movq 0x38(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x1648f74
movq 0x38(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x1648f72
jmp 0x1648fa1
movq 0x38(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x2a0(%rsp)
cmpq $0x0, 0x2a0(%rsp)
je 0x1648f9f
movq 0x2a0(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x1648fa1
jmp 0x1648fa3
movq 0x38(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x1648ffb
movq %rax, %rdi
callq 0x678a0
jmp 0x1648ffd
leaq 0xc8(%rsp), %rax
movq %rax, 0x160(%rsp)
movq 0x160(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x28(%rsp)
leaq 0xc8(%rsp), %rax
movq %rax, 0x170(%rsp)
movq 0x170(%rsp), %rax
movq %rax, 0x288(%rsp)
movq 0x288(%rsp), %rax
movq %rax, 0x30(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x16490dc
movq 0x30(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x284(%rsp) # imm = 0xFFFFFFFF
movl 0x284(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x280(%rsp)
cmpl $0x1, 0x280(%rsp)
jne 0x16490dc
movq 0x30(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x16490ad
movq 0x30(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x16490ab
jmp 0x16490da
movq 0x30(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x2a8(%rsp)
cmpq $0x0, 0x2a8(%rsp)
je 0x16490d8
movq 0x2a8(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x16490da
jmp 0x16490dc
movq 0x30(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x1649134
movq %rax, %rdi
callq 0x678a0
movq 0x28(%rsp), %rax
movq %rax, 0x110(%rsp)
movl $0x0, 0xc4(%rsp)
movl 0xc4(%rsp), %eax
addl $0xf, %eax
cmpl 0x11c(%rsp), %eax
jge 0x1649206
movq 0x110(%rsp), %rax
movq %rax, 0x178(%rsp)
movq 0x178(%rsp), %rax
vmovups (%rax), %zmm0
vmovaps %zmm0, 0x80(%rsp)
leaq 0x137(%rsp), %rdi
leaq 0x80(%rsp), %rsi
callq 0x164e090
vmovaps %zmm0, 0x80(%rsp)
movq 0x110(%rsp), %rax
vmovaps 0x80(%rsp), %zmm0
movq %rax, 0x1d0(%rsp)
vmovaps %zmm0, 0x180(%rsp)
vmovaps 0x180(%rsp), %zmm0
movq 0x1d0(%rsp), %rax
vmovups %zmm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x40, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x10, %eax
movl %eax, 0xc4(%rsp)
jmp 0x164914c
jmp 0x1649208
movl 0xc4(%rsp), %eax
addl $0x7, %eax
cmpl 0x11c(%rsp), %eax
jge 0x16492b7
movq 0x110(%rsp), %rax
movq %rax, 0x1d8(%rsp)
movq 0x1d8(%rsp), %rax
vmovups (%rax), %ymm0
vmovaps %ymm0, 0x60(%rsp)
leaq 0x137(%rsp), %rdi
leaq 0x60(%rsp), %rsi
callq 0x164e6f0
vmovaps %ymm0, 0x60(%rsp)
movq 0x110(%rsp), %rax
vmovaps 0x60(%rsp), %ymm0
movq %rax, 0x210(%rsp)
vmovaps %ymm0, 0x1e0(%rsp)
vmovaps 0x1e0(%rsp), %ymm0
movq 0x210(%rsp), %rax
vmovups %ymm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x20, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x8, %eax
movl %eax, 0xc4(%rsp)
jmp 0x1649208
jmp 0x16492b9
movl 0xc4(%rsp), %eax
addl $0x3, %eax
cmpl 0x11c(%rsp), %eax
jge 0x164936b
movq 0x110(%rsp), %rax
movq %rax, 0x218(%rsp)
movq 0x218(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm0, 0x50(%rsp)
leaq 0x137(%rsp), %rdi
leaq 0x50(%rsp), %rsi
vzeroupper
callq 0x164ef70
vmovaps %xmm0, 0x50(%rsp)
movq 0x110(%rsp), %rax
vmovaps 0x50(%rsp), %xmm0
movq %rax, 0x230(%rsp)
vmovaps %xmm0, 0x220(%rsp)
vmovaps 0x220(%rsp), %xmm0
movq 0x230(%rsp), %rax
vmovaps %xmm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x10, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x4, %eax
movl %eax, 0xc4(%rsp)
jmp 0x16492b9
jmp 0x164936d
movl 0xc4(%rsp), %eax
cmpl 0x11c(%rsp), %eax
jge 0x16493c8
movq 0x110(%rsp), %rsi
leaq 0x137(%rsp), %rdi
vzeroupper
callq 0x164f730
movq 0x110(%rsp), %rax
vmovss %xmm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x4, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x1, %eax
movl %eax, 0xc4(%rsp)
jmp 0x164936d
jmp 0x16493ca
movl 0x118(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x118(%rsp)
jmp 0x1648d0a
xorl %eax, %eax
movq %rbp, %rsp
popq %rbp
vzeroupper
retq
nopw (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,243 |
int ncnn::unary_op_inplace<ncnn::UnaryOp_x86_avx512_functor::unary_op_log>(ncnn::Mat&, ncnn::Option const&)
|
static int unary_op_inplace(Mat& a, const Option& opt)
{
Op op;
int w = a.w;
int h = a.h;
int d = a.d;
int channels = a.c;
int elempack = a.elempack;
int size = w * h * d * elempack;
#pragma omp parallel for num_threads(opt.num_threads)
for (int q = 0; q < channels; q++)
{
float* ptr = a.channel(q);
int i = 0;
#if __SSE2__
#if __AVX__
#if __AVX512F__
for (; i + 15 < size; i += 16)
{
__m512 _p = _mm512_loadu_ps(ptr);
_p = op.func_pack16(_p);
_mm512_storeu_ps(ptr, _p);
ptr += 16;
}
#endif // __AVX512F__
for (; i + 7 < size; i += 8)
{
__m256 _p = _mm256_loadu_ps(ptr);
_p = op.func_pack8(_p);
_mm256_storeu_ps(ptr, _p);
ptr += 8;
}
#endif // __AVX__
for (; i + 3 < size; i += 4)
{
__m128 _p = _mm_load_ps(ptr);
_p = op.func_pack4(_p);
_mm_store_ps(ptr, _p);
ptr += 4;
}
#endif // __SSE2__
for (; i < size; i++)
{
*ptr = op.func(*ptr);
ptr++;
}
}
return 0;
}
|
pushq %rbp
movq %rsp, %rbp
andq $-0x40, %rsp
subq $0x2c0, %rsp # imm = 0x2C0
movq %rdi, 0x140(%rsp)
movq %rsi, 0x138(%rsp)
movq 0x140(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x130(%rsp)
movq 0x140(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0x12c(%rsp)
movq 0x140(%rsp), %rax
movl 0x34(%rax), %eax
movl %eax, 0x128(%rsp)
movq 0x140(%rsp), %rax
movl 0x38(%rax), %eax
movl %eax, 0x124(%rsp)
movq 0x140(%rsp), %rax
movl 0x18(%rax), %eax
movl %eax, 0x120(%rsp)
movl 0x130(%rsp), %eax
imull 0x12c(%rsp), %eax
imull 0x128(%rsp), %eax
imull 0x120(%rsp), %eax
movl %eax, 0x11c(%rsp)
movl $0x0, 0x118(%rsp)
movl 0x118(%rsp), %eax
cmpl 0x124(%rsp), %eax
jge 0x1649b70
movq 0x140(%rsp), %rcx
movl 0x118(%rsp), %eax
leaq 0xc8(%rsp), %rdx
movq %rdx, 0x158(%rsp)
movq %rcx, 0x150(%rsp)
movl %eax, 0x14c(%rsp)
movq 0x150(%rsp), %rax
movq %rax, 0x48(%rsp)
movb $0x0, 0x14b(%rsp)
movl 0x2c(%rax), %r9d
movl 0x30(%rax), %r8d
movl 0x34(%rax), %edi
movq (%rax), %rsi
movq 0x40(%rax), %rcx
movslq 0x14c(%rsp), %rdx
imulq %rdx, %rcx
imulq 0x10(%rax), %rcx
addq %rcx, %rsi
movq 0x10(%rax), %rdx
movl 0x18(%rax), %ecx
movq 0x20(%rax), %rax
leaq 0xc8(%rsp), %r10
movq %r10, 0x268(%rsp)
movl %r9d, 0x264(%rsp)
movl %r8d, 0x260(%rsp)
movl %edi, 0x25c(%rsp)
movq %rsi, 0x250(%rsp)
movq %rdx, 0x248(%rsp)
movl %ecx, 0x244(%rsp)
movq %rax, 0x238(%rsp)
movq 0x268(%rsp), %rcx
movq %rcx, 0x40(%rsp)
movq 0x250(%rsp), %rax
movq %rax, (%rcx)
movq $0x0, 0x8(%rcx)
movq 0x248(%rsp), %rax
movq %rax, 0x10(%rcx)
movl 0x244(%rsp), %eax
movl %eax, 0x18(%rcx)
movq 0x238(%rsp), %rax
movq %rax, 0x20(%rcx)
movl $0x3, 0x28(%rcx)
movl 0x264(%rsp), %eax
movl %eax, 0x2c(%rcx)
movl 0x260(%rsp), %eax
movl %eax, 0x30(%rcx)
movl $0x1, 0x34(%rcx)
movl 0x25c(%rsp), %eax
movl %eax, 0x38(%rcx)
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rdx
imulq %rdx, %rax
imulq 0x10(%rcx), %rax
movq %rax, 0x278(%rsp)
movl $0x10, 0x274(%rsp)
movq 0x278(%rsp), %rax
movslq 0x274(%rsp), %rdx
addq %rdx, %rax
subq $0x1, %rax
xorl %edx, %edx
subl 0x274(%rsp), %edx
movslq %edx, %rdx
andq %rdx, %rax
xorl %edx, %edx
divq 0x10(%rcx)
movq %rax, %rdx
movq 0x48(%rsp), %rax
movq %rdx, 0x40(%rcx)
movl 0x28(%rax), %ecx
subl $0x1, %ecx
movl %ecx, 0xf0(%rsp)
cmpl $0x4, 0x28(%rax)
jne 0x164965e
movq 0x48(%rsp), %rcx
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rcx
imulq %rcx, %rax
movq %rax, 0x108(%rsp)
movb $0x1, 0x14b(%rsp)
testb $0x1, 0x14b(%rsp)
jne 0x164978d
leaq 0xc8(%rsp), %rax
movq %rax, 0x168(%rsp)
movq 0x168(%rsp), %rax
movq %rax, 0x298(%rsp)
movq 0x298(%rsp), %rax
movq %rax, 0x38(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x1649733
movq 0x38(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x294(%rsp) # imm = 0xFFFFFFFF
movl 0x294(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x290(%rsp)
cmpl $0x1, 0x290(%rsp)
jne 0x1649733
movq 0x38(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x1649704
movq 0x38(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x1649702
jmp 0x1649731
movq 0x38(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x2a0(%rsp)
cmpq $0x0, 0x2a0(%rsp)
je 0x164972f
movq 0x2a0(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x1649731
jmp 0x1649733
movq 0x38(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x164978b
movq %rax, %rdi
callq 0x678a0
jmp 0x164978d
leaq 0xc8(%rsp), %rax
movq %rax, 0x160(%rsp)
movq 0x160(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x28(%rsp)
leaq 0xc8(%rsp), %rax
movq %rax, 0x170(%rsp)
movq 0x170(%rsp), %rax
movq %rax, 0x288(%rsp)
movq 0x288(%rsp), %rax
movq %rax, 0x30(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x164986c
movq 0x30(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x284(%rsp) # imm = 0xFFFFFFFF
movl 0x284(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x280(%rsp)
cmpl $0x1, 0x280(%rsp)
jne 0x164986c
movq 0x30(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x164983d
movq 0x30(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x164983b
jmp 0x164986a
movq 0x30(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x2a8(%rsp)
cmpq $0x0, 0x2a8(%rsp)
je 0x1649868
movq 0x2a8(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x164986a
jmp 0x164986c
movq 0x30(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x16498c4
movq %rax, %rdi
callq 0x678a0
movq 0x28(%rsp), %rax
movq %rax, 0x110(%rsp)
movl $0x0, 0xc4(%rsp)
movl 0xc4(%rsp), %eax
addl $0xf, %eax
cmpl 0x11c(%rsp), %eax
jge 0x1649996
movq 0x110(%rsp), %rax
movq %rax, 0x178(%rsp)
movq 0x178(%rsp), %rax
vmovups (%rax), %zmm0
vmovaps %zmm0, 0x80(%rsp)
leaq 0x137(%rsp), %rdi
leaq 0x80(%rsp), %rsi
callq 0x164f760
vmovaps %zmm0, 0x80(%rsp)
movq 0x110(%rsp), %rax
vmovaps 0x80(%rsp), %zmm0
movq %rax, 0x1d0(%rsp)
vmovaps %zmm0, 0x180(%rsp)
vmovaps 0x180(%rsp), %zmm0
movq 0x1d0(%rsp), %rax
vmovups %zmm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x40, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x10, %eax
movl %eax, 0xc4(%rsp)
jmp 0x16498dc
jmp 0x1649998
movl 0xc4(%rsp), %eax
addl $0x7, %eax
cmpl 0x11c(%rsp), %eax
jge 0x1649a47
movq 0x110(%rsp), %rax
movq %rax, 0x1d8(%rsp)
movq 0x1d8(%rsp), %rax
vmovups (%rax), %ymm0
vmovaps %ymm0, 0x60(%rsp)
leaq 0x137(%rsp), %rdi
leaq 0x60(%rsp), %rsi
callq 0x1650030
vmovaps %ymm0, 0x60(%rsp)
movq 0x110(%rsp), %rax
vmovaps 0x60(%rsp), %ymm0
movq %rax, 0x210(%rsp)
vmovaps %ymm0, 0x1e0(%rsp)
vmovaps 0x1e0(%rsp), %ymm0
movq 0x210(%rsp), %rax
vmovups %ymm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x20, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x8, %eax
movl %eax, 0xc4(%rsp)
jmp 0x1649998
jmp 0x1649a49
movl 0xc4(%rsp), %eax
addl $0x3, %eax
cmpl 0x11c(%rsp), %eax
jge 0x1649afb
movq 0x110(%rsp), %rax
movq %rax, 0x218(%rsp)
movq 0x218(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm0, 0x50(%rsp)
leaq 0x137(%rsp), %rdi
leaq 0x50(%rsp), %rsi
vzeroupper
callq 0x1650b60
vmovaps %xmm0, 0x50(%rsp)
movq 0x110(%rsp), %rax
vmovaps 0x50(%rsp), %xmm0
movq %rax, 0x230(%rsp)
vmovaps %xmm0, 0x220(%rsp)
vmovaps 0x220(%rsp), %xmm0
movq 0x230(%rsp), %rax
vmovaps %xmm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x10, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x4, %eax
movl %eax, 0xc4(%rsp)
jmp 0x1649a49
jmp 0x1649afd
movl 0xc4(%rsp), %eax
cmpl 0x11c(%rsp), %eax
jge 0x1649b58
movq 0x110(%rsp), %rsi
leaq 0x137(%rsp), %rdi
vzeroupper
callq 0x16515d0
movq 0x110(%rsp), %rax
vmovss %xmm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x4, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x1, %eax
movl %eax, 0xc4(%rsp)
jmp 0x1649afd
jmp 0x1649b5a
movl 0x118(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x118(%rsp)
jmp 0x164949a
xorl %eax, %eax
movq %rbp, %rsp
popq %rbp
vzeroupper
retq
nopw (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,244 |
int ncnn::unary_op_inplace<ncnn::UnaryOp_x86_avx512_functor::unary_op_sin>(ncnn::Mat&, ncnn::Option const&)
|
static int unary_op_inplace(Mat& a, const Option& opt)
{
Op op;
int w = a.w;
int h = a.h;
int d = a.d;
int channels = a.c;
int elempack = a.elempack;
int size = w * h * d * elempack;
#pragma omp parallel for num_threads(opt.num_threads)
for (int q = 0; q < channels; q++)
{
float* ptr = a.channel(q);
int i = 0;
#if __SSE2__
#if __AVX__
#if __AVX512F__
for (; i + 15 < size; i += 16)
{
__m512 _p = _mm512_loadu_ps(ptr);
_p = op.func_pack16(_p);
_mm512_storeu_ps(ptr, _p);
ptr += 16;
}
#endif // __AVX512F__
for (; i + 7 < size; i += 8)
{
__m256 _p = _mm256_loadu_ps(ptr);
_p = op.func_pack8(_p);
_mm256_storeu_ps(ptr, _p);
ptr += 8;
}
#endif // __AVX__
for (; i + 3 < size; i += 4)
{
__m128 _p = _mm_load_ps(ptr);
_p = op.func_pack4(_p);
_mm_store_ps(ptr, _p);
ptr += 4;
}
#endif // __SSE2__
for (; i < size; i++)
{
*ptr = op.func(*ptr);
ptr++;
}
}
return 0;
}
|
pushq %rbp
movq %rsp, %rbp
andq $-0x40, %rsp
subq $0x2c0, %rsp # imm = 0x2C0
movq %rdi, 0x140(%rsp)
movq %rsi, 0x138(%rsp)
movq 0x140(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x130(%rsp)
movq 0x140(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0x12c(%rsp)
movq 0x140(%rsp), %rax
movl 0x34(%rax), %eax
movl %eax, 0x128(%rsp)
movq 0x140(%rsp), %rax
movl 0x38(%rax), %eax
movl %eax, 0x124(%rsp)
movq 0x140(%rsp), %rax
movl 0x18(%rax), %eax
movl %eax, 0x120(%rsp)
movl 0x130(%rsp), %eax
imull 0x12c(%rsp), %eax
imull 0x128(%rsp), %eax
imull 0x120(%rsp), %eax
movl %eax, 0x11c(%rsp)
movl $0x0, 0x118(%rsp)
movl 0x118(%rsp), %eax
cmpl 0x124(%rsp), %eax
jge 0x164a300
movq 0x140(%rsp), %rcx
movl 0x118(%rsp), %eax
leaq 0xc8(%rsp), %rdx
movq %rdx, 0x158(%rsp)
movq %rcx, 0x150(%rsp)
movl %eax, 0x14c(%rsp)
movq 0x150(%rsp), %rax
movq %rax, 0x48(%rsp)
movb $0x0, 0x14b(%rsp)
movl 0x2c(%rax), %r9d
movl 0x30(%rax), %r8d
movl 0x34(%rax), %edi
movq (%rax), %rsi
movq 0x40(%rax), %rcx
movslq 0x14c(%rsp), %rdx
imulq %rdx, %rcx
imulq 0x10(%rax), %rcx
addq %rcx, %rsi
movq 0x10(%rax), %rdx
movl 0x18(%rax), %ecx
movq 0x20(%rax), %rax
leaq 0xc8(%rsp), %r10
movq %r10, 0x268(%rsp)
movl %r9d, 0x264(%rsp)
movl %r8d, 0x260(%rsp)
movl %edi, 0x25c(%rsp)
movq %rsi, 0x250(%rsp)
movq %rdx, 0x248(%rsp)
movl %ecx, 0x244(%rsp)
movq %rax, 0x238(%rsp)
movq 0x268(%rsp), %rcx
movq %rcx, 0x40(%rsp)
movq 0x250(%rsp), %rax
movq %rax, (%rcx)
movq $0x0, 0x8(%rcx)
movq 0x248(%rsp), %rax
movq %rax, 0x10(%rcx)
movl 0x244(%rsp), %eax
movl %eax, 0x18(%rcx)
movq 0x238(%rsp), %rax
movq %rax, 0x20(%rcx)
movl $0x3, 0x28(%rcx)
movl 0x264(%rsp), %eax
movl %eax, 0x2c(%rcx)
movl 0x260(%rsp), %eax
movl %eax, 0x30(%rcx)
movl $0x1, 0x34(%rcx)
movl 0x25c(%rsp), %eax
movl %eax, 0x38(%rcx)
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rdx
imulq %rdx, %rax
imulq 0x10(%rcx), %rax
movq %rax, 0x278(%rsp)
movl $0x10, 0x274(%rsp)
movq 0x278(%rsp), %rax
movslq 0x274(%rsp), %rdx
addq %rdx, %rax
subq $0x1, %rax
xorl %edx, %edx
subl 0x274(%rsp), %edx
movslq %edx, %rdx
andq %rdx, %rax
xorl %edx, %edx
divq 0x10(%rcx)
movq %rax, %rdx
movq 0x48(%rsp), %rax
movq %rdx, 0x40(%rcx)
movl 0x28(%rax), %ecx
subl $0x1, %ecx
movl %ecx, 0xf0(%rsp)
cmpl $0x4, 0x28(%rax)
jne 0x1649dee
movq 0x48(%rsp), %rcx
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rcx
imulq %rcx, %rax
movq %rax, 0x108(%rsp)
movb $0x1, 0x14b(%rsp)
testb $0x1, 0x14b(%rsp)
jne 0x1649f1d
leaq 0xc8(%rsp), %rax
movq %rax, 0x168(%rsp)
movq 0x168(%rsp), %rax
movq %rax, 0x298(%rsp)
movq 0x298(%rsp), %rax
movq %rax, 0x38(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x1649ec3
movq 0x38(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x294(%rsp) # imm = 0xFFFFFFFF
movl 0x294(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x290(%rsp)
cmpl $0x1, 0x290(%rsp)
jne 0x1649ec3
movq 0x38(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x1649e94
movq 0x38(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x1649e92
jmp 0x1649ec1
movq 0x38(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x2a0(%rsp)
cmpq $0x0, 0x2a0(%rsp)
je 0x1649ebf
movq 0x2a0(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x1649ec1
jmp 0x1649ec3
movq 0x38(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x1649f1b
movq %rax, %rdi
callq 0x678a0
jmp 0x1649f1d
leaq 0xc8(%rsp), %rax
movq %rax, 0x160(%rsp)
movq 0x160(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x28(%rsp)
leaq 0xc8(%rsp), %rax
movq %rax, 0x170(%rsp)
movq 0x170(%rsp), %rax
movq %rax, 0x288(%rsp)
movq 0x288(%rsp), %rax
movq %rax, 0x30(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x1649ffc
movq 0x30(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x284(%rsp) # imm = 0xFFFFFFFF
movl 0x284(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x280(%rsp)
cmpl $0x1, 0x280(%rsp)
jne 0x1649ffc
movq 0x30(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x1649fcd
movq 0x30(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x1649fcb
jmp 0x1649ffa
movq 0x30(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x2a8(%rsp)
cmpq $0x0, 0x2a8(%rsp)
je 0x1649ff8
movq 0x2a8(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x1649ffa
jmp 0x1649ffc
movq 0x30(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x164a054
movq %rax, %rdi
callq 0x678a0
movq 0x28(%rsp), %rax
movq %rax, 0x110(%rsp)
movl $0x0, 0xc4(%rsp)
movl 0xc4(%rsp), %eax
addl $0xf, %eax
cmpl 0x11c(%rsp), %eax
jge 0x164a126
movq 0x110(%rsp), %rax
movq %rax, 0x178(%rsp)
movq 0x178(%rsp), %rax
vmovups (%rax), %zmm0
vmovaps %zmm0, 0x80(%rsp)
leaq 0x137(%rsp), %rdi
leaq 0x80(%rsp), %rsi
callq 0x1651600
vmovaps %zmm0, 0x80(%rsp)
movq 0x110(%rsp), %rax
vmovaps 0x80(%rsp), %zmm0
movq %rax, 0x1d0(%rsp)
vmovaps %zmm0, 0x180(%rsp)
vmovaps 0x180(%rsp), %zmm0
movq 0x1d0(%rsp), %rax
vmovups %zmm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x40, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x10, %eax
movl %eax, 0xc4(%rsp)
jmp 0x164a06c
jmp 0x164a128
movl 0xc4(%rsp), %eax
addl $0x7, %eax
cmpl 0x11c(%rsp), %eax
jge 0x164a1d7
movq 0x110(%rsp), %rax
movq %rax, 0x1d8(%rsp)
movq 0x1d8(%rsp), %rax
vmovups (%rax), %ymm0
vmovaps %ymm0, 0x60(%rsp)
leaq 0x137(%rsp), %rdi
leaq 0x60(%rsp), %rsi
callq 0x1651f20
vmovaps %ymm0, 0x60(%rsp)
movq 0x110(%rsp), %rax
vmovaps 0x60(%rsp), %ymm0
movq %rax, 0x210(%rsp)
vmovaps %ymm0, 0x1e0(%rsp)
vmovaps 0x1e0(%rsp), %ymm0
movq 0x210(%rsp), %rax
vmovups %ymm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x20, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x8, %eax
movl %eax, 0xc4(%rsp)
jmp 0x164a128
jmp 0x164a1d9
movl 0xc4(%rsp), %eax
addl $0x3, %eax
cmpl 0x11c(%rsp), %eax
jge 0x164a28b
movq 0x110(%rsp), %rax
movq %rax, 0x218(%rsp)
movq 0x218(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm0, 0x50(%rsp)
leaq 0x137(%rsp), %rdi
leaq 0x50(%rsp), %rsi
vzeroupper
callq 0x1652aa0
vmovaps %xmm0, 0x50(%rsp)
movq 0x110(%rsp), %rax
vmovaps 0x50(%rsp), %xmm0
movq %rax, 0x230(%rsp)
vmovaps %xmm0, 0x220(%rsp)
vmovaps 0x220(%rsp), %xmm0
movq 0x230(%rsp), %rax
vmovaps %xmm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x10, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x4, %eax
movl %eax, 0xc4(%rsp)
jmp 0x164a1d9
jmp 0x164a28d
movl 0xc4(%rsp), %eax
cmpl 0x11c(%rsp), %eax
jge 0x164a2e8
movq 0x110(%rsp), %rsi
leaq 0x137(%rsp), %rdi
vzeroupper
callq 0x1653500
movq 0x110(%rsp), %rax
vmovss %xmm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x4, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x1, %eax
movl %eax, 0xc4(%rsp)
jmp 0x164a28d
jmp 0x164a2ea
movl 0x118(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x118(%rsp)
jmp 0x1649c2a
xorl %eax, %eax
movq %rbp, %rsp
popq %rbp
vzeroupper
retq
nopw (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,245 |
int ncnn::unary_op_inplace<ncnn::UnaryOp_x86_avx512_functor::unary_op_cos>(ncnn::Mat&, ncnn::Option const&)
|
static int unary_op_inplace(Mat& a, const Option& opt)
{
Op op;
int w = a.w;
int h = a.h;
int d = a.d;
int channels = a.c;
int elempack = a.elempack;
int size = w * h * d * elempack;
#pragma omp parallel for num_threads(opt.num_threads)
for (int q = 0; q < channels; q++)
{
float* ptr = a.channel(q);
int i = 0;
#if __SSE2__
#if __AVX__
#if __AVX512F__
for (; i + 15 < size; i += 16)
{
__m512 _p = _mm512_loadu_ps(ptr);
_p = op.func_pack16(_p);
_mm512_storeu_ps(ptr, _p);
ptr += 16;
}
#endif // __AVX512F__
for (; i + 7 < size; i += 8)
{
__m256 _p = _mm256_loadu_ps(ptr);
_p = op.func_pack8(_p);
_mm256_storeu_ps(ptr, _p);
ptr += 8;
}
#endif // __AVX__
for (; i + 3 < size; i += 4)
{
__m128 _p = _mm_load_ps(ptr);
_p = op.func_pack4(_p);
_mm_store_ps(ptr, _p);
ptr += 4;
}
#endif // __SSE2__
for (; i < size; i++)
{
*ptr = op.func(*ptr);
ptr++;
}
}
return 0;
}
|
pushq %rbp
movq %rsp, %rbp
andq $-0x40, %rsp
subq $0x2c0, %rsp # imm = 0x2C0
movq %rdi, 0x140(%rsp)
movq %rsi, 0x138(%rsp)
movq 0x140(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x130(%rsp)
movq 0x140(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0x12c(%rsp)
movq 0x140(%rsp), %rax
movl 0x34(%rax), %eax
movl %eax, 0x128(%rsp)
movq 0x140(%rsp), %rax
movl 0x38(%rax), %eax
movl %eax, 0x124(%rsp)
movq 0x140(%rsp), %rax
movl 0x18(%rax), %eax
movl %eax, 0x120(%rsp)
movl 0x130(%rsp), %eax
imull 0x12c(%rsp), %eax
imull 0x128(%rsp), %eax
imull 0x120(%rsp), %eax
movl %eax, 0x11c(%rsp)
movl $0x0, 0x118(%rsp)
movl 0x118(%rsp), %eax
cmpl 0x124(%rsp), %eax
jge 0x164aa90
movq 0x140(%rsp), %rcx
movl 0x118(%rsp), %eax
leaq 0xc8(%rsp), %rdx
movq %rdx, 0x158(%rsp)
movq %rcx, 0x150(%rsp)
movl %eax, 0x14c(%rsp)
movq 0x150(%rsp), %rax
movq %rax, 0x48(%rsp)
movb $0x0, 0x14b(%rsp)
movl 0x2c(%rax), %r9d
movl 0x30(%rax), %r8d
movl 0x34(%rax), %edi
movq (%rax), %rsi
movq 0x40(%rax), %rcx
movslq 0x14c(%rsp), %rdx
imulq %rdx, %rcx
imulq 0x10(%rax), %rcx
addq %rcx, %rsi
movq 0x10(%rax), %rdx
movl 0x18(%rax), %ecx
movq 0x20(%rax), %rax
leaq 0xc8(%rsp), %r10
movq %r10, 0x268(%rsp)
movl %r9d, 0x264(%rsp)
movl %r8d, 0x260(%rsp)
movl %edi, 0x25c(%rsp)
movq %rsi, 0x250(%rsp)
movq %rdx, 0x248(%rsp)
movl %ecx, 0x244(%rsp)
movq %rax, 0x238(%rsp)
movq 0x268(%rsp), %rcx
movq %rcx, 0x40(%rsp)
movq 0x250(%rsp), %rax
movq %rax, (%rcx)
movq $0x0, 0x8(%rcx)
movq 0x248(%rsp), %rax
movq %rax, 0x10(%rcx)
movl 0x244(%rsp), %eax
movl %eax, 0x18(%rcx)
movq 0x238(%rsp), %rax
movq %rax, 0x20(%rcx)
movl $0x3, 0x28(%rcx)
movl 0x264(%rsp), %eax
movl %eax, 0x2c(%rcx)
movl 0x260(%rsp), %eax
movl %eax, 0x30(%rcx)
movl $0x1, 0x34(%rcx)
movl 0x25c(%rsp), %eax
movl %eax, 0x38(%rcx)
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rdx
imulq %rdx, %rax
imulq 0x10(%rcx), %rax
movq %rax, 0x278(%rsp)
movl $0x10, 0x274(%rsp)
movq 0x278(%rsp), %rax
movslq 0x274(%rsp), %rdx
addq %rdx, %rax
subq $0x1, %rax
xorl %edx, %edx
subl 0x274(%rsp), %edx
movslq %edx, %rdx
andq %rdx, %rax
xorl %edx, %edx
divq 0x10(%rcx)
movq %rax, %rdx
movq 0x48(%rsp), %rax
movq %rdx, 0x40(%rcx)
movl 0x28(%rax), %ecx
subl $0x1, %ecx
movl %ecx, 0xf0(%rsp)
cmpl $0x4, 0x28(%rax)
jne 0x164a57e
movq 0x48(%rsp), %rcx
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rcx
imulq %rcx, %rax
movq %rax, 0x108(%rsp)
movb $0x1, 0x14b(%rsp)
testb $0x1, 0x14b(%rsp)
jne 0x164a6ad
leaq 0xc8(%rsp), %rax
movq %rax, 0x168(%rsp)
movq 0x168(%rsp), %rax
movq %rax, 0x298(%rsp)
movq 0x298(%rsp), %rax
movq %rax, 0x38(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x164a653
movq 0x38(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x294(%rsp) # imm = 0xFFFFFFFF
movl 0x294(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x290(%rsp)
cmpl $0x1, 0x290(%rsp)
jne 0x164a653
movq 0x38(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x164a624
movq 0x38(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x164a622
jmp 0x164a651
movq 0x38(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x2a0(%rsp)
cmpq $0x0, 0x2a0(%rsp)
je 0x164a64f
movq 0x2a0(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x164a651
jmp 0x164a653
movq 0x38(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x164a6ab
movq %rax, %rdi
callq 0x678a0
jmp 0x164a6ad
leaq 0xc8(%rsp), %rax
movq %rax, 0x160(%rsp)
movq 0x160(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x28(%rsp)
leaq 0xc8(%rsp), %rax
movq %rax, 0x170(%rsp)
movq 0x170(%rsp), %rax
movq %rax, 0x288(%rsp)
movq 0x288(%rsp), %rax
movq %rax, 0x30(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x164a78c
movq 0x30(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x284(%rsp) # imm = 0xFFFFFFFF
movl 0x284(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x280(%rsp)
cmpl $0x1, 0x280(%rsp)
jne 0x164a78c
movq 0x30(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x164a75d
movq 0x30(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x164a75b
jmp 0x164a78a
movq 0x30(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x2a8(%rsp)
cmpq $0x0, 0x2a8(%rsp)
je 0x164a788
movq 0x2a8(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x164a78a
jmp 0x164a78c
movq 0x30(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x164a7e4
movq %rax, %rdi
callq 0x678a0
movq 0x28(%rsp), %rax
movq %rax, 0x110(%rsp)
movl $0x0, 0xc4(%rsp)
movl 0xc4(%rsp), %eax
addl $0xf, %eax
cmpl 0x11c(%rsp), %eax
jge 0x164a8b6
movq 0x110(%rsp), %rax
movq %rax, 0x178(%rsp)
movq 0x178(%rsp), %rax
vmovups (%rax), %zmm0
vmovaps %zmm0, 0x80(%rsp)
leaq 0x137(%rsp), %rdi
leaq 0x80(%rsp), %rsi
callq 0x1653530
vmovaps %zmm0, 0x80(%rsp)
movq 0x110(%rsp), %rax
vmovaps 0x80(%rsp), %zmm0
movq %rax, 0x1d0(%rsp)
vmovaps %zmm0, 0x180(%rsp)
vmovaps 0x180(%rsp), %zmm0
movq 0x1d0(%rsp), %rax
vmovups %zmm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x40, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x10, %eax
movl %eax, 0xc4(%rsp)
jmp 0x164a7fc
jmp 0x164a8b8
movl 0xc4(%rsp), %eax
addl $0x7, %eax
cmpl 0x11c(%rsp), %eax
jge 0x164a967
movq 0x110(%rsp), %rax
movq %rax, 0x1d8(%rsp)
movq 0x1d8(%rsp), %rax
vmovups (%rax), %ymm0
vmovaps %ymm0, 0x60(%rsp)
leaq 0x137(%rsp), %rdi
leaq 0x60(%rsp), %rsi
callq 0x1653e70
vmovaps %ymm0, 0x60(%rsp)
movq 0x110(%rsp), %rax
vmovaps 0x60(%rsp), %ymm0
movq %rax, 0x210(%rsp)
vmovaps %ymm0, 0x1e0(%rsp)
vmovaps 0x1e0(%rsp), %ymm0
movq 0x210(%rsp), %rax
vmovups %ymm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x20, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x8, %eax
movl %eax, 0xc4(%rsp)
jmp 0x164a8b8
jmp 0x164a969
movl 0xc4(%rsp), %eax
addl $0x3, %eax
cmpl 0x11c(%rsp), %eax
jge 0x164aa1b
movq 0x110(%rsp), %rax
movq %rax, 0x218(%rsp)
movq 0x218(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm0, 0x50(%rsp)
leaq 0x137(%rsp), %rdi
leaq 0x50(%rsp), %rsi
vzeroupper
callq 0x1654980
vmovaps %xmm0, 0x50(%rsp)
movq 0x110(%rsp), %rax
vmovaps 0x50(%rsp), %xmm0
movq %rax, 0x230(%rsp)
vmovaps %xmm0, 0x220(%rsp)
vmovaps 0x220(%rsp), %xmm0
movq 0x230(%rsp), %rax
vmovaps %xmm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x10, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x4, %eax
movl %eax, 0xc4(%rsp)
jmp 0x164a969
jmp 0x164aa1d
movl 0xc4(%rsp), %eax
cmpl 0x11c(%rsp), %eax
jge 0x164aa78
movq 0x110(%rsp), %rsi
leaq 0x137(%rsp), %rdi
vzeroupper
callq 0x1655390
movq 0x110(%rsp), %rax
vmovss %xmm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x4, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x1, %eax
movl %eax, 0xc4(%rsp)
jmp 0x164aa1d
jmp 0x164aa7a
movl 0x118(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x118(%rsp)
jmp 0x164a3ba
xorl %eax, %eax
movq %rbp, %rsp
popq %rbp
vzeroupper
retq
nopw (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,246 |
int ncnn::unary_op_inplace<ncnn::UnaryOp_x86_avx512_functor::unary_op_tan>(ncnn::Mat&, ncnn::Option const&)
|
static int unary_op_inplace(Mat& a, const Option& opt)
{
Op op;
int w = a.w;
int h = a.h;
int d = a.d;
int channels = a.c;
int elempack = a.elempack;
int size = w * h * d * elempack;
#pragma omp parallel for num_threads(opt.num_threads)
for (int q = 0; q < channels; q++)
{
float* ptr = a.channel(q);
int i = 0;
#if __SSE2__
#if __AVX__
#if __AVX512F__
for (; i + 15 < size; i += 16)
{
__m512 _p = _mm512_loadu_ps(ptr);
_p = op.func_pack16(_p);
_mm512_storeu_ps(ptr, _p);
ptr += 16;
}
#endif // __AVX512F__
for (; i + 7 < size; i += 8)
{
__m256 _p = _mm256_loadu_ps(ptr);
_p = op.func_pack8(_p);
_mm256_storeu_ps(ptr, _p);
ptr += 8;
}
#endif // __AVX__
for (; i + 3 < size; i += 4)
{
__m128 _p = _mm_load_ps(ptr);
_p = op.func_pack4(_p);
_mm_store_ps(ptr, _p);
ptr += 4;
}
#endif // __SSE2__
for (; i < size; i++)
{
*ptr = op.func(*ptr);
ptr++;
}
}
return 0;
}
|
pushq %rbp
movq %rsp, %rbp
andq $-0x40, %rsp
subq $0x2c0, %rsp # imm = 0x2C0
movq %rdi, 0x140(%rsp)
movq %rsi, 0x138(%rsp)
movq 0x140(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x130(%rsp)
movq 0x140(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0x12c(%rsp)
movq 0x140(%rsp), %rax
movl 0x34(%rax), %eax
movl %eax, 0x128(%rsp)
movq 0x140(%rsp), %rax
movl 0x38(%rax), %eax
movl %eax, 0x124(%rsp)
movq 0x140(%rsp), %rax
movl 0x18(%rax), %eax
movl %eax, 0x120(%rsp)
movl 0x130(%rsp), %eax
imull 0x12c(%rsp), %eax
imull 0x128(%rsp), %eax
imull 0x120(%rsp), %eax
movl %eax, 0x11c(%rsp)
movl $0x0, 0x118(%rsp)
movl 0x118(%rsp), %eax
cmpl 0x124(%rsp), %eax
jge 0x164b220
movq 0x140(%rsp), %rcx
movl 0x118(%rsp), %eax
leaq 0xc8(%rsp), %rdx
movq %rdx, 0x158(%rsp)
movq %rcx, 0x150(%rsp)
movl %eax, 0x14c(%rsp)
movq 0x150(%rsp), %rax
movq %rax, 0x48(%rsp)
movb $0x0, 0x14b(%rsp)
movl 0x2c(%rax), %r9d
movl 0x30(%rax), %r8d
movl 0x34(%rax), %edi
movq (%rax), %rsi
movq 0x40(%rax), %rcx
movslq 0x14c(%rsp), %rdx
imulq %rdx, %rcx
imulq 0x10(%rax), %rcx
addq %rcx, %rsi
movq 0x10(%rax), %rdx
movl 0x18(%rax), %ecx
movq 0x20(%rax), %rax
leaq 0xc8(%rsp), %r10
movq %r10, 0x268(%rsp)
movl %r9d, 0x264(%rsp)
movl %r8d, 0x260(%rsp)
movl %edi, 0x25c(%rsp)
movq %rsi, 0x250(%rsp)
movq %rdx, 0x248(%rsp)
movl %ecx, 0x244(%rsp)
movq %rax, 0x238(%rsp)
movq 0x268(%rsp), %rcx
movq %rcx, 0x40(%rsp)
movq 0x250(%rsp), %rax
movq %rax, (%rcx)
movq $0x0, 0x8(%rcx)
movq 0x248(%rsp), %rax
movq %rax, 0x10(%rcx)
movl 0x244(%rsp), %eax
movl %eax, 0x18(%rcx)
movq 0x238(%rsp), %rax
movq %rax, 0x20(%rcx)
movl $0x3, 0x28(%rcx)
movl 0x264(%rsp), %eax
movl %eax, 0x2c(%rcx)
movl 0x260(%rsp), %eax
movl %eax, 0x30(%rcx)
movl $0x1, 0x34(%rcx)
movl 0x25c(%rsp), %eax
movl %eax, 0x38(%rcx)
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rdx
imulq %rdx, %rax
imulq 0x10(%rcx), %rax
movq %rax, 0x278(%rsp)
movl $0x10, 0x274(%rsp)
movq 0x278(%rsp), %rax
movslq 0x274(%rsp), %rdx
addq %rdx, %rax
subq $0x1, %rax
xorl %edx, %edx
subl 0x274(%rsp), %edx
movslq %edx, %rdx
andq %rdx, %rax
xorl %edx, %edx
divq 0x10(%rcx)
movq %rax, %rdx
movq 0x48(%rsp), %rax
movq %rdx, 0x40(%rcx)
movl 0x28(%rax), %ecx
subl $0x1, %ecx
movl %ecx, 0xf0(%rsp)
cmpl $0x4, 0x28(%rax)
jne 0x164ad0e
movq 0x48(%rsp), %rcx
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rcx
imulq %rcx, %rax
movq %rax, 0x108(%rsp)
movb $0x1, 0x14b(%rsp)
testb $0x1, 0x14b(%rsp)
jne 0x164ae3d
leaq 0xc8(%rsp), %rax
movq %rax, 0x168(%rsp)
movq 0x168(%rsp), %rax
movq %rax, 0x298(%rsp)
movq 0x298(%rsp), %rax
movq %rax, 0x38(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x164ade3
movq 0x38(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x294(%rsp) # imm = 0xFFFFFFFF
movl 0x294(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x290(%rsp)
cmpl $0x1, 0x290(%rsp)
jne 0x164ade3
movq 0x38(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x164adb4
movq 0x38(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x164adb2
jmp 0x164ade1
movq 0x38(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x2a0(%rsp)
cmpq $0x0, 0x2a0(%rsp)
je 0x164addf
movq 0x2a0(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x164ade1
jmp 0x164ade3
movq 0x38(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x164ae3b
movq %rax, %rdi
callq 0x678a0
jmp 0x164ae3d
leaq 0xc8(%rsp), %rax
movq %rax, 0x160(%rsp)
movq 0x160(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x28(%rsp)
leaq 0xc8(%rsp), %rax
movq %rax, 0x170(%rsp)
movq 0x170(%rsp), %rax
movq %rax, 0x288(%rsp)
movq 0x288(%rsp), %rax
movq %rax, 0x30(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x164af1c
movq 0x30(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x284(%rsp) # imm = 0xFFFFFFFF
movl 0x284(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x280(%rsp)
cmpl $0x1, 0x280(%rsp)
jne 0x164af1c
movq 0x30(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x164aeed
movq 0x30(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x164aeeb
jmp 0x164af1a
movq 0x30(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x2a8(%rsp)
cmpq $0x0, 0x2a8(%rsp)
je 0x164af18
movq 0x2a8(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x164af1a
jmp 0x164af1c
movq 0x30(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x164af74
movq %rax, %rdi
callq 0x678a0
movq 0x28(%rsp), %rax
movq %rax, 0x110(%rsp)
movl $0x0, 0xc4(%rsp)
movl 0xc4(%rsp), %eax
addl $0xf, %eax
cmpl 0x11c(%rsp), %eax
jge 0x164b046
movq 0x110(%rsp), %rax
movq %rax, 0x178(%rsp)
movq 0x178(%rsp), %rax
vmovups (%rax), %zmm0
vmovaps %zmm0, 0x80(%rsp)
leaq 0x137(%rsp), %rdi
leaq 0x80(%rsp), %rsi
callq 0x16553c0
vmovaps %zmm0, 0x80(%rsp)
movq 0x110(%rsp), %rax
vmovaps 0x80(%rsp), %zmm0
movq %rax, 0x1d0(%rsp)
vmovaps %zmm0, 0x180(%rsp)
vmovaps 0x180(%rsp), %zmm0
movq 0x1d0(%rsp), %rax
vmovups %zmm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x40, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x10, %eax
movl %eax, 0xc4(%rsp)
jmp 0x164af8c
jmp 0x164b048
movl 0xc4(%rsp), %eax
addl $0x7, %eax
cmpl 0x11c(%rsp), %eax
jge 0x164b0f7
movq 0x110(%rsp), %rax
movq %rax, 0x1d8(%rsp)
movq 0x1d8(%rsp), %rax
vmovups (%rax), %ymm0
vmovaps %ymm0, 0x60(%rsp)
leaq 0x137(%rsp), %rdi
leaq 0x60(%rsp), %rsi
callq 0x1656010
vmovaps %ymm0, 0x60(%rsp)
movq 0x110(%rsp), %rax
vmovaps 0x60(%rsp), %ymm0
movq %rax, 0x210(%rsp)
vmovaps %ymm0, 0x1e0(%rsp)
vmovaps 0x1e0(%rsp), %ymm0
movq 0x210(%rsp), %rax
vmovups %ymm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x20, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x8, %eax
movl %eax, 0xc4(%rsp)
jmp 0x164b048
jmp 0x164b0f9
movl 0xc4(%rsp), %eax
addl $0x3, %eax
cmpl 0x11c(%rsp), %eax
jge 0x164b1ab
movq 0x110(%rsp), %rax
movq %rax, 0x218(%rsp)
movq 0x218(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm0, 0x50(%rsp)
leaq 0x137(%rsp), %rdi
leaq 0x50(%rsp), %rsi
vzeroupper
callq 0x1657010
vmovaps %xmm0, 0x50(%rsp)
movq 0x110(%rsp), %rax
vmovaps 0x50(%rsp), %xmm0
movq %rax, 0x230(%rsp)
vmovaps %xmm0, 0x220(%rsp)
vmovaps 0x220(%rsp), %xmm0
movq 0x230(%rsp), %rax
vmovaps %xmm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x10, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x4, %eax
movl %eax, 0xc4(%rsp)
jmp 0x164b0f9
jmp 0x164b1ad
movl 0xc4(%rsp), %eax
cmpl 0x11c(%rsp), %eax
jge 0x164b208
movq 0x110(%rsp), %rsi
leaq 0x137(%rsp), %rdi
vzeroupper
callq 0x1657eb0
movq 0x110(%rsp), %rax
vmovss %xmm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x4, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x1, %eax
movl %eax, 0xc4(%rsp)
jmp 0x164b1ad
jmp 0x164b20a
movl 0x118(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x118(%rsp)
jmp 0x164ab4a
xorl %eax, %eax
movq %rbp, %rsp
popq %rbp
vzeroupper
retq
nopw (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,247 |
int ncnn::unary_op_inplace<ncnn::UnaryOp_x86_avx512_functor::unary_op_asin>(ncnn::Mat&, ncnn::Option const&)
|
static int unary_op_inplace(Mat& a, const Option& opt)
{
Op op;
int w = a.w;
int h = a.h;
int d = a.d;
int channels = a.c;
int elempack = a.elempack;
int size = w * h * d * elempack;
#pragma omp parallel for num_threads(opt.num_threads)
for (int q = 0; q < channels; q++)
{
float* ptr = a.channel(q);
int i = 0;
#if __SSE2__
#if __AVX__
#if __AVX512F__
for (; i + 15 < size; i += 16)
{
__m512 _p = _mm512_loadu_ps(ptr);
_p = op.func_pack16(_p);
_mm512_storeu_ps(ptr, _p);
ptr += 16;
}
#endif // __AVX512F__
for (; i + 7 < size; i += 8)
{
__m256 _p = _mm256_loadu_ps(ptr);
_p = op.func_pack8(_p);
_mm256_storeu_ps(ptr, _p);
ptr += 8;
}
#endif // __AVX__
for (; i + 3 < size; i += 4)
{
__m128 _p = _mm_load_ps(ptr);
_p = op.func_pack4(_p);
_mm_store_ps(ptr, _p);
ptr += 4;
}
#endif // __SSE2__
for (; i < size; i++)
{
*ptr = op.func(*ptr);
ptr++;
}
}
return 0;
}
|
pushq %rbp
movq %rsp, %rbp
andq $-0x40, %rsp
subq $0x2c0, %rsp # imm = 0x2C0
movq %rdi, 0x140(%rsp)
movq %rsi, 0x138(%rsp)
movq 0x140(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x130(%rsp)
movq 0x140(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0x12c(%rsp)
movq 0x140(%rsp), %rax
movl 0x34(%rax), %eax
movl %eax, 0x128(%rsp)
movq 0x140(%rsp), %rax
movl 0x38(%rax), %eax
movl %eax, 0x124(%rsp)
movq 0x140(%rsp), %rax
movl 0x18(%rax), %eax
movl %eax, 0x120(%rsp)
movl 0x130(%rsp), %eax
imull 0x12c(%rsp), %eax
imull 0x128(%rsp), %eax
imull 0x120(%rsp), %eax
movl %eax, 0x11c(%rsp)
movl $0x0, 0x118(%rsp)
movl 0x118(%rsp), %eax
cmpl 0x124(%rsp), %eax
jge 0x164b9b0
movq 0x140(%rsp), %rcx
movl 0x118(%rsp), %eax
leaq 0xc8(%rsp), %rdx
movq %rdx, 0x158(%rsp)
movq %rcx, 0x150(%rsp)
movl %eax, 0x14c(%rsp)
movq 0x150(%rsp), %rax
movq %rax, 0x48(%rsp)
movb $0x0, 0x14b(%rsp)
movl 0x2c(%rax), %r9d
movl 0x30(%rax), %r8d
movl 0x34(%rax), %edi
movq (%rax), %rsi
movq 0x40(%rax), %rcx
movslq 0x14c(%rsp), %rdx
imulq %rdx, %rcx
imulq 0x10(%rax), %rcx
addq %rcx, %rsi
movq 0x10(%rax), %rdx
movl 0x18(%rax), %ecx
movq 0x20(%rax), %rax
leaq 0xc8(%rsp), %r10
movq %r10, 0x268(%rsp)
movl %r9d, 0x264(%rsp)
movl %r8d, 0x260(%rsp)
movl %edi, 0x25c(%rsp)
movq %rsi, 0x250(%rsp)
movq %rdx, 0x248(%rsp)
movl %ecx, 0x244(%rsp)
movq %rax, 0x238(%rsp)
movq 0x268(%rsp), %rcx
movq %rcx, 0x40(%rsp)
movq 0x250(%rsp), %rax
movq %rax, (%rcx)
movq $0x0, 0x8(%rcx)
movq 0x248(%rsp), %rax
movq %rax, 0x10(%rcx)
movl 0x244(%rsp), %eax
movl %eax, 0x18(%rcx)
movq 0x238(%rsp), %rax
movq %rax, 0x20(%rcx)
movl $0x3, 0x28(%rcx)
movl 0x264(%rsp), %eax
movl %eax, 0x2c(%rcx)
movl 0x260(%rsp), %eax
movl %eax, 0x30(%rcx)
movl $0x1, 0x34(%rcx)
movl 0x25c(%rsp), %eax
movl %eax, 0x38(%rcx)
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rdx
imulq %rdx, %rax
imulq 0x10(%rcx), %rax
movq %rax, 0x278(%rsp)
movl $0x10, 0x274(%rsp)
movq 0x278(%rsp), %rax
movslq 0x274(%rsp), %rdx
addq %rdx, %rax
subq $0x1, %rax
xorl %edx, %edx
subl 0x274(%rsp), %edx
movslq %edx, %rdx
andq %rdx, %rax
xorl %edx, %edx
divq 0x10(%rcx)
movq %rax, %rdx
movq 0x48(%rsp), %rax
movq %rdx, 0x40(%rcx)
movl 0x28(%rax), %ecx
subl $0x1, %ecx
movl %ecx, 0xf0(%rsp)
cmpl $0x4, 0x28(%rax)
jne 0x164b49e
movq 0x48(%rsp), %rcx
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rcx
imulq %rcx, %rax
movq %rax, 0x108(%rsp)
movb $0x1, 0x14b(%rsp)
testb $0x1, 0x14b(%rsp)
jne 0x164b5cd
leaq 0xc8(%rsp), %rax
movq %rax, 0x168(%rsp)
movq 0x168(%rsp), %rax
movq %rax, 0x298(%rsp)
movq 0x298(%rsp), %rax
movq %rax, 0x38(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x164b573
movq 0x38(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x294(%rsp) # imm = 0xFFFFFFFF
movl 0x294(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x290(%rsp)
cmpl $0x1, 0x290(%rsp)
jne 0x164b573
movq 0x38(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x164b544
movq 0x38(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x164b542
jmp 0x164b571
movq 0x38(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x2a0(%rsp)
cmpq $0x0, 0x2a0(%rsp)
je 0x164b56f
movq 0x2a0(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x164b571
jmp 0x164b573
movq 0x38(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x164b5cb
movq %rax, %rdi
callq 0x678a0
jmp 0x164b5cd
leaq 0xc8(%rsp), %rax
movq %rax, 0x160(%rsp)
movq 0x160(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x28(%rsp)
leaq 0xc8(%rsp), %rax
movq %rax, 0x170(%rsp)
movq 0x170(%rsp), %rax
movq %rax, 0x288(%rsp)
movq 0x288(%rsp), %rax
movq %rax, 0x30(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x164b6ac
movq 0x30(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x284(%rsp) # imm = 0xFFFFFFFF
movl 0x284(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x280(%rsp)
cmpl $0x1, 0x280(%rsp)
jne 0x164b6ac
movq 0x30(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x164b67d
movq 0x30(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x164b67b
jmp 0x164b6aa
movq 0x30(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x2a8(%rsp)
cmpq $0x0, 0x2a8(%rsp)
je 0x164b6a8
movq 0x2a8(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x164b6aa
jmp 0x164b6ac
movq 0x30(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x164b704
movq %rax, %rdi
callq 0x678a0
movq 0x28(%rsp), %rax
movq %rax, 0x110(%rsp)
movl $0x0, 0xc4(%rsp)
movl 0xc4(%rsp), %eax
addl $0xf, %eax
cmpl 0x11c(%rsp), %eax
jge 0x164b7d6
movq 0x110(%rsp), %rax
movq %rax, 0x178(%rsp)
movq 0x178(%rsp), %rax
vmovups (%rax), %zmm0
vmovaps %zmm0, 0x80(%rsp)
leaq 0x137(%rsp), %rdi
leaq 0x80(%rsp), %rsi
callq 0x1657ee0
vmovaps %zmm0, 0x80(%rsp)
movq 0x110(%rsp), %rax
vmovaps 0x80(%rsp), %zmm0
movq %rax, 0x1d0(%rsp)
vmovaps %zmm0, 0x180(%rsp)
vmovaps 0x180(%rsp), %zmm0
movq 0x1d0(%rsp), %rax
vmovups %zmm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x40, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x10, %eax
movl %eax, 0xc4(%rsp)
jmp 0x164b71c
jmp 0x164b7d8
movl 0xc4(%rsp), %eax
addl $0x7, %eax
cmpl 0x11c(%rsp), %eax
jge 0x164b887
movq 0x110(%rsp), %rax
movq %rax, 0x1d8(%rsp)
movq 0x1d8(%rsp), %rax
vmovups (%rax), %ymm0
vmovaps %ymm0, 0x60(%rsp)
leaq 0x137(%rsp), %rdi
leaq 0x60(%rsp), %rsi
callq 0x1657f90
vmovaps %ymm0, 0x60(%rsp)
movq 0x110(%rsp), %rax
vmovaps 0x60(%rsp), %ymm0
movq %rax, 0x210(%rsp)
vmovaps %ymm0, 0x1e0(%rsp)
vmovaps 0x1e0(%rsp), %ymm0
movq 0x210(%rsp), %rax
vmovups %ymm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x20, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x8, %eax
movl %eax, 0xc4(%rsp)
jmp 0x164b7d8
jmp 0x164b889
movl 0xc4(%rsp), %eax
addl $0x3, %eax
cmpl 0x11c(%rsp), %eax
jge 0x164b93b
movq 0x110(%rsp), %rax
movq %rax, 0x218(%rsp)
movq 0x218(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm0, 0x50(%rsp)
leaq 0x137(%rsp), %rdi
leaq 0x50(%rsp), %rsi
vzeroupper
callq 0x1658070
vmovaps %xmm0, 0x50(%rsp)
movq 0x110(%rsp), %rax
vmovaps 0x50(%rsp), %xmm0
movq %rax, 0x230(%rsp)
vmovaps %xmm0, 0x220(%rsp)
vmovaps 0x220(%rsp), %xmm0
movq 0x230(%rsp), %rax
vmovaps %xmm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x10, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x4, %eax
movl %eax, 0xc4(%rsp)
jmp 0x164b889
jmp 0x164b93d
movl 0xc4(%rsp), %eax
cmpl 0x11c(%rsp), %eax
jge 0x164b998
movq 0x110(%rsp), %rsi
leaq 0x137(%rsp), %rdi
vzeroupper
callq 0x1658100
movq 0x110(%rsp), %rax
vmovss %xmm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x4, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x1, %eax
movl %eax, 0xc4(%rsp)
jmp 0x164b93d
jmp 0x164b99a
movl 0x118(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x118(%rsp)
jmp 0x164b2da
xorl %eax, %eax
movq %rbp, %rsp
popq %rbp
vzeroupper
retq
nopw (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,248 |
int ncnn::unary_op_inplace<ncnn::UnaryOp_x86_avx512_functor::unary_op_acos>(ncnn::Mat&, ncnn::Option const&)
|
static int unary_op_inplace(Mat& a, const Option& opt)
{
Op op;
int w = a.w;
int h = a.h;
int d = a.d;
int channels = a.c;
int elempack = a.elempack;
int size = w * h * d * elempack;
#pragma omp parallel for num_threads(opt.num_threads)
for (int q = 0; q < channels; q++)
{
float* ptr = a.channel(q);
int i = 0;
#if __SSE2__
#if __AVX__
#if __AVX512F__
for (; i + 15 < size; i += 16)
{
__m512 _p = _mm512_loadu_ps(ptr);
_p = op.func_pack16(_p);
_mm512_storeu_ps(ptr, _p);
ptr += 16;
}
#endif // __AVX512F__
for (; i + 7 < size; i += 8)
{
__m256 _p = _mm256_loadu_ps(ptr);
_p = op.func_pack8(_p);
_mm256_storeu_ps(ptr, _p);
ptr += 8;
}
#endif // __AVX__
for (; i + 3 < size; i += 4)
{
__m128 _p = _mm_load_ps(ptr);
_p = op.func_pack4(_p);
_mm_store_ps(ptr, _p);
ptr += 4;
}
#endif // __SSE2__
for (; i < size; i++)
{
*ptr = op.func(*ptr);
ptr++;
}
}
return 0;
}
|
pushq %rbp
movq %rsp, %rbp
andq $-0x40, %rsp
subq $0x2c0, %rsp # imm = 0x2C0
movq %rdi, 0x140(%rsp)
movq %rsi, 0x138(%rsp)
movq 0x140(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x130(%rsp)
movq 0x140(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0x12c(%rsp)
movq 0x140(%rsp), %rax
movl 0x34(%rax), %eax
movl %eax, 0x128(%rsp)
movq 0x140(%rsp), %rax
movl 0x38(%rax), %eax
movl %eax, 0x124(%rsp)
movq 0x140(%rsp), %rax
movl 0x18(%rax), %eax
movl %eax, 0x120(%rsp)
movl 0x130(%rsp), %eax
imull 0x12c(%rsp), %eax
imull 0x128(%rsp), %eax
imull 0x120(%rsp), %eax
movl %eax, 0x11c(%rsp)
movl $0x0, 0x118(%rsp)
movl 0x118(%rsp), %eax
cmpl 0x124(%rsp), %eax
jge 0x164c140
movq 0x140(%rsp), %rcx
movl 0x118(%rsp), %eax
leaq 0xc8(%rsp), %rdx
movq %rdx, 0x158(%rsp)
movq %rcx, 0x150(%rsp)
movl %eax, 0x14c(%rsp)
movq 0x150(%rsp), %rax
movq %rax, 0x48(%rsp)
movb $0x0, 0x14b(%rsp)
movl 0x2c(%rax), %r9d
movl 0x30(%rax), %r8d
movl 0x34(%rax), %edi
movq (%rax), %rsi
movq 0x40(%rax), %rcx
movslq 0x14c(%rsp), %rdx
imulq %rdx, %rcx
imulq 0x10(%rax), %rcx
addq %rcx, %rsi
movq 0x10(%rax), %rdx
movl 0x18(%rax), %ecx
movq 0x20(%rax), %rax
leaq 0xc8(%rsp), %r10
movq %r10, 0x268(%rsp)
movl %r9d, 0x264(%rsp)
movl %r8d, 0x260(%rsp)
movl %edi, 0x25c(%rsp)
movq %rsi, 0x250(%rsp)
movq %rdx, 0x248(%rsp)
movl %ecx, 0x244(%rsp)
movq %rax, 0x238(%rsp)
movq 0x268(%rsp), %rcx
movq %rcx, 0x40(%rsp)
movq 0x250(%rsp), %rax
movq %rax, (%rcx)
movq $0x0, 0x8(%rcx)
movq 0x248(%rsp), %rax
movq %rax, 0x10(%rcx)
movl 0x244(%rsp), %eax
movl %eax, 0x18(%rcx)
movq 0x238(%rsp), %rax
movq %rax, 0x20(%rcx)
movl $0x3, 0x28(%rcx)
movl 0x264(%rsp), %eax
movl %eax, 0x2c(%rcx)
movl 0x260(%rsp), %eax
movl %eax, 0x30(%rcx)
movl $0x1, 0x34(%rcx)
movl 0x25c(%rsp), %eax
movl %eax, 0x38(%rcx)
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rdx
imulq %rdx, %rax
imulq 0x10(%rcx), %rax
movq %rax, 0x278(%rsp)
movl $0x10, 0x274(%rsp)
movq 0x278(%rsp), %rax
movslq 0x274(%rsp), %rdx
addq %rdx, %rax
subq $0x1, %rax
xorl %edx, %edx
subl 0x274(%rsp), %edx
movslq %edx, %rdx
andq %rdx, %rax
xorl %edx, %edx
divq 0x10(%rcx)
movq %rax, %rdx
movq 0x48(%rsp), %rax
movq %rdx, 0x40(%rcx)
movl 0x28(%rax), %ecx
subl $0x1, %ecx
movl %ecx, 0xf0(%rsp)
cmpl $0x4, 0x28(%rax)
jne 0x164bc2e
movq 0x48(%rsp), %rcx
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rcx
imulq %rcx, %rax
movq %rax, 0x108(%rsp)
movb $0x1, 0x14b(%rsp)
testb $0x1, 0x14b(%rsp)
jne 0x164bd5d
leaq 0xc8(%rsp), %rax
movq %rax, 0x168(%rsp)
movq 0x168(%rsp), %rax
movq %rax, 0x298(%rsp)
movq 0x298(%rsp), %rax
movq %rax, 0x38(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x164bd03
movq 0x38(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x294(%rsp) # imm = 0xFFFFFFFF
movl 0x294(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x290(%rsp)
cmpl $0x1, 0x290(%rsp)
jne 0x164bd03
movq 0x38(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x164bcd4
movq 0x38(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x164bcd2
jmp 0x164bd01
movq 0x38(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x2a0(%rsp)
cmpq $0x0, 0x2a0(%rsp)
je 0x164bcff
movq 0x2a0(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x164bd01
jmp 0x164bd03
movq 0x38(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x164bd5b
movq %rax, %rdi
callq 0x678a0
jmp 0x164bd5d
leaq 0xc8(%rsp), %rax
movq %rax, 0x160(%rsp)
movq 0x160(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x28(%rsp)
leaq 0xc8(%rsp), %rax
movq %rax, 0x170(%rsp)
movq 0x170(%rsp), %rax
movq %rax, 0x288(%rsp)
movq 0x288(%rsp), %rax
movq %rax, 0x30(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x164be3c
movq 0x30(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x284(%rsp) # imm = 0xFFFFFFFF
movl 0x284(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x280(%rsp)
cmpl $0x1, 0x280(%rsp)
jne 0x164be3c
movq 0x30(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x164be0d
movq 0x30(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x164be0b
jmp 0x164be3a
movq 0x30(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x2a8(%rsp)
cmpq $0x0, 0x2a8(%rsp)
je 0x164be38
movq 0x2a8(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x164be3a
jmp 0x164be3c
movq 0x30(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x164be94
movq %rax, %rdi
callq 0x678a0
movq 0x28(%rsp), %rax
movq %rax, 0x110(%rsp)
movl $0x0, 0xc4(%rsp)
movl 0xc4(%rsp), %eax
addl $0xf, %eax
cmpl 0x11c(%rsp), %eax
jge 0x164bf66
movq 0x110(%rsp), %rax
movq %rax, 0x178(%rsp)
movq 0x178(%rsp), %rax
vmovups (%rax), %zmm0
vmovaps %zmm0, 0x80(%rsp)
leaq 0x137(%rsp), %rdi
leaq 0x80(%rsp), %rsi
callq 0x1658130
vmovaps %zmm0, 0x80(%rsp)
movq 0x110(%rsp), %rax
vmovaps 0x80(%rsp), %zmm0
movq %rax, 0x1d0(%rsp)
vmovaps %zmm0, 0x180(%rsp)
vmovaps 0x180(%rsp), %zmm0
movq 0x1d0(%rsp), %rax
vmovups %zmm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x40, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x10, %eax
movl %eax, 0xc4(%rsp)
jmp 0x164beac
jmp 0x164bf68
movl 0xc4(%rsp), %eax
addl $0x7, %eax
cmpl 0x11c(%rsp), %eax
jge 0x164c017
movq 0x110(%rsp), %rax
movq %rax, 0x1d8(%rsp)
movq 0x1d8(%rsp), %rax
vmovups (%rax), %ymm0
vmovaps %ymm0, 0x60(%rsp)
leaq 0x137(%rsp), %rdi
leaq 0x60(%rsp), %rsi
callq 0x16581e0
vmovaps %ymm0, 0x60(%rsp)
movq 0x110(%rsp), %rax
vmovaps 0x60(%rsp), %ymm0
movq %rax, 0x210(%rsp)
vmovaps %ymm0, 0x1e0(%rsp)
vmovaps 0x1e0(%rsp), %ymm0
movq 0x210(%rsp), %rax
vmovups %ymm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x20, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x8, %eax
movl %eax, 0xc4(%rsp)
jmp 0x164bf68
jmp 0x164c019
movl 0xc4(%rsp), %eax
addl $0x3, %eax
cmpl 0x11c(%rsp), %eax
jge 0x164c0cb
movq 0x110(%rsp), %rax
movq %rax, 0x218(%rsp)
movq 0x218(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm0, 0x50(%rsp)
leaq 0x137(%rsp), %rdi
leaq 0x50(%rsp), %rsi
vzeroupper
callq 0x16582c0
vmovaps %xmm0, 0x50(%rsp)
movq 0x110(%rsp), %rax
vmovaps 0x50(%rsp), %xmm0
movq %rax, 0x230(%rsp)
vmovaps %xmm0, 0x220(%rsp)
vmovaps 0x220(%rsp), %xmm0
movq 0x230(%rsp), %rax
vmovaps %xmm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x10, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x4, %eax
movl %eax, 0xc4(%rsp)
jmp 0x164c019
jmp 0x164c0cd
movl 0xc4(%rsp), %eax
cmpl 0x11c(%rsp), %eax
jge 0x164c128
movq 0x110(%rsp), %rsi
leaq 0x137(%rsp), %rdi
vzeroupper
callq 0x1658350
movq 0x110(%rsp), %rax
vmovss %xmm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x4, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x1, %eax
movl %eax, 0xc4(%rsp)
jmp 0x164c0cd
jmp 0x164c12a
movl 0x118(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x118(%rsp)
jmp 0x164ba6a
xorl %eax, %eax
movq %rbp, %rsp
popq %rbp
vzeroupper
retq
nopw (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,249 |
int ncnn::unary_op_inplace<ncnn::UnaryOp_x86_avx512_functor::unary_op_atan>(ncnn::Mat&, ncnn::Option const&)
|
static int unary_op_inplace(Mat& a, const Option& opt)
{
Op op;
int w = a.w;
int h = a.h;
int d = a.d;
int channels = a.c;
int elempack = a.elempack;
int size = w * h * d * elempack;
#pragma omp parallel for num_threads(opt.num_threads)
for (int q = 0; q < channels; q++)
{
float* ptr = a.channel(q);
int i = 0;
#if __SSE2__
#if __AVX__
#if __AVX512F__
for (; i + 15 < size; i += 16)
{
__m512 _p = _mm512_loadu_ps(ptr);
_p = op.func_pack16(_p);
_mm512_storeu_ps(ptr, _p);
ptr += 16;
}
#endif // __AVX512F__
for (; i + 7 < size; i += 8)
{
__m256 _p = _mm256_loadu_ps(ptr);
_p = op.func_pack8(_p);
_mm256_storeu_ps(ptr, _p);
ptr += 8;
}
#endif // __AVX__
for (; i + 3 < size; i += 4)
{
__m128 _p = _mm_load_ps(ptr);
_p = op.func_pack4(_p);
_mm_store_ps(ptr, _p);
ptr += 4;
}
#endif // __SSE2__
for (; i < size; i++)
{
*ptr = op.func(*ptr);
ptr++;
}
}
return 0;
}
|
pushq %rbp
movq %rsp, %rbp
andq $-0x40, %rsp
subq $0x2c0, %rsp # imm = 0x2C0
movq %rdi, 0x140(%rsp)
movq %rsi, 0x138(%rsp)
movq 0x140(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x130(%rsp)
movq 0x140(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0x12c(%rsp)
movq 0x140(%rsp), %rax
movl 0x34(%rax), %eax
movl %eax, 0x128(%rsp)
movq 0x140(%rsp), %rax
movl 0x38(%rax), %eax
movl %eax, 0x124(%rsp)
movq 0x140(%rsp), %rax
movl 0x18(%rax), %eax
movl %eax, 0x120(%rsp)
movl 0x130(%rsp), %eax
imull 0x12c(%rsp), %eax
imull 0x128(%rsp), %eax
imull 0x120(%rsp), %eax
movl %eax, 0x11c(%rsp)
movl $0x0, 0x118(%rsp)
movl 0x118(%rsp), %eax
cmpl 0x124(%rsp), %eax
jge 0x164c8d0
movq 0x140(%rsp), %rcx
movl 0x118(%rsp), %eax
leaq 0xc8(%rsp), %rdx
movq %rdx, 0x158(%rsp)
movq %rcx, 0x150(%rsp)
movl %eax, 0x14c(%rsp)
movq 0x150(%rsp), %rax
movq %rax, 0x48(%rsp)
movb $0x0, 0x14b(%rsp)
movl 0x2c(%rax), %r9d
movl 0x30(%rax), %r8d
movl 0x34(%rax), %edi
movq (%rax), %rsi
movq 0x40(%rax), %rcx
movslq 0x14c(%rsp), %rdx
imulq %rdx, %rcx
imulq 0x10(%rax), %rcx
addq %rcx, %rsi
movq 0x10(%rax), %rdx
movl 0x18(%rax), %ecx
movq 0x20(%rax), %rax
leaq 0xc8(%rsp), %r10
movq %r10, 0x268(%rsp)
movl %r9d, 0x264(%rsp)
movl %r8d, 0x260(%rsp)
movl %edi, 0x25c(%rsp)
movq %rsi, 0x250(%rsp)
movq %rdx, 0x248(%rsp)
movl %ecx, 0x244(%rsp)
movq %rax, 0x238(%rsp)
movq 0x268(%rsp), %rcx
movq %rcx, 0x40(%rsp)
movq 0x250(%rsp), %rax
movq %rax, (%rcx)
movq $0x0, 0x8(%rcx)
movq 0x248(%rsp), %rax
movq %rax, 0x10(%rcx)
movl 0x244(%rsp), %eax
movl %eax, 0x18(%rcx)
movq 0x238(%rsp), %rax
movq %rax, 0x20(%rcx)
movl $0x3, 0x28(%rcx)
movl 0x264(%rsp), %eax
movl %eax, 0x2c(%rcx)
movl 0x260(%rsp), %eax
movl %eax, 0x30(%rcx)
movl $0x1, 0x34(%rcx)
movl 0x25c(%rsp), %eax
movl %eax, 0x38(%rcx)
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rdx
imulq %rdx, %rax
imulq 0x10(%rcx), %rax
movq %rax, 0x278(%rsp)
movl $0x10, 0x274(%rsp)
movq 0x278(%rsp), %rax
movslq 0x274(%rsp), %rdx
addq %rdx, %rax
subq $0x1, %rax
xorl %edx, %edx
subl 0x274(%rsp), %edx
movslq %edx, %rdx
andq %rdx, %rax
xorl %edx, %edx
divq 0x10(%rcx)
movq %rax, %rdx
movq 0x48(%rsp), %rax
movq %rdx, 0x40(%rcx)
movl 0x28(%rax), %ecx
subl $0x1, %ecx
movl %ecx, 0xf0(%rsp)
cmpl $0x4, 0x28(%rax)
jne 0x164c3be
movq 0x48(%rsp), %rcx
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rcx
imulq %rcx, %rax
movq %rax, 0x108(%rsp)
movb $0x1, 0x14b(%rsp)
testb $0x1, 0x14b(%rsp)
jne 0x164c4ed
leaq 0xc8(%rsp), %rax
movq %rax, 0x168(%rsp)
movq 0x168(%rsp), %rax
movq %rax, 0x298(%rsp)
movq 0x298(%rsp), %rax
movq %rax, 0x38(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x164c493
movq 0x38(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x294(%rsp) # imm = 0xFFFFFFFF
movl 0x294(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x290(%rsp)
cmpl $0x1, 0x290(%rsp)
jne 0x164c493
movq 0x38(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x164c464
movq 0x38(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x164c462
jmp 0x164c491
movq 0x38(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x2a0(%rsp)
cmpq $0x0, 0x2a0(%rsp)
je 0x164c48f
movq 0x2a0(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x164c491
jmp 0x164c493
movq 0x38(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x164c4eb
movq %rax, %rdi
callq 0x678a0
jmp 0x164c4ed
leaq 0xc8(%rsp), %rax
movq %rax, 0x160(%rsp)
movq 0x160(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x28(%rsp)
leaq 0xc8(%rsp), %rax
movq %rax, 0x170(%rsp)
movq 0x170(%rsp), %rax
movq %rax, 0x288(%rsp)
movq 0x288(%rsp), %rax
movq %rax, 0x30(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x164c5cc
movq 0x30(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x284(%rsp) # imm = 0xFFFFFFFF
movl 0x284(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x280(%rsp)
cmpl $0x1, 0x280(%rsp)
jne 0x164c5cc
movq 0x30(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x164c59d
movq 0x30(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x164c59b
jmp 0x164c5ca
movq 0x30(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x2a8(%rsp)
cmpq $0x0, 0x2a8(%rsp)
je 0x164c5c8
movq 0x2a8(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x164c5ca
jmp 0x164c5cc
movq 0x30(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x164c624
movq %rax, %rdi
callq 0x678a0
movq 0x28(%rsp), %rax
movq %rax, 0x110(%rsp)
movl $0x0, 0xc4(%rsp)
movl 0xc4(%rsp), %eax
addl $0xf, %eax
cmpl 0x11c(%rsp), %eax
jge 0x164c6f6
movq 0x110(%rsp), %rax
movq %rax, 0x178(%rsp)
movq 0x178(%rsp), %rax
vmovups (%rax), %zmm0
vmovaps %zmm0, 0x80(%rsp)
leaq 0x137(%rsp), %rdi
leaq 0x80(%rsp), %rsi
callq 0x1658380
vmovaps %zmm0, 0x80(%rsp)
movq 0x110(%rsp), %rax
vmovaps 0x80(%rsp), %zmm0
movq %rax, 0x1d0(%rsp)
vmovaps %zmm0, 0x180(%rsp)
vmovaps 0x180(%rsp), %zmm0
movq 0x1d0(%rsp), %rax
vmovups %zmm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x40, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x10, %eax
movl %eax, 0xc4(%rsp)
jmp 0x164c63c
jmp 0x164c6f8
movl 0xc4(%rsp), %eax
addl $0x7, %eax
cmpl 0x11c(%rsp), %eax
jge 0x164c7a7
movq 0x110(%rsp), %rax
movq %rax, 0x1d8(%rsp)
movq 0x1d8(%rsp), %rax
vmovups (%rax), %ymm0
vmovaps %ymm0, 0x60(%rsp)
leaq 0x137(%rsp), %rdi
leaq 0x60(%rsp), %rsi
callq 0x1658430
vmovaps %ymm0, 0x60(%rsp)
movq 0x110(%rsp), %rax
vmovaps 0x60(%rsp), %ymm0
movq %rax, 0x210(%rsp)
vmovaps %ymm0, 0x1e0(%rsp)
vmovaps 0x1e0(%rsp), %ymm0
movq 0x210(%rsp), %rax
vmovups %ymm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x20, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x8, %eax
movl %eax, 0xc4(%rsp)
jmp 0x164c6f8
jmp 0x164c7a9
movl 0xc4(%rsp), %eax
addl $0x3, %eax
cmpl 0x11c(%rsp), %eax
jge 0x164c85b
movq 0x110(%rsp), %rax
movq %rax, 0x218(%rsp)
movq 0x218(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm0, 0x50(%rsp)
leaq 0x137(%rsp), %rdi
leaq 0x50(%rsp), %rsi
vzeroupper
callq 0x1658510
vmovaps %xmm0, 0x50(%rsp)
movq 0x110(%rsp), %rax
vmovaps 0x50(%rsp), %xmm0
movq %rax, 0x230(%rsp)
vmovaps %xmm0, 0x220(%rsp)
vmovaps 0x220(%rsp), %xmm0
movq 0x230(%rsp), %rax
vmovaps %xmm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x10, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x4, %eax
movl %eax, 0xc4(%rsp)
jmp 0x164c7a9
jmp 0x164c85d
movl 0xc4(%rsp), %eax
cmpl 0x11c(%rsp), %eax
jge 0x164c8b8
movq 0x110(%rsp), %rsi
leaq 0x137(%rsp), %rdi
vzeroupper
callq 0x16585a0
movq 0x110(%rsp), %rax
vmovss %xmm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x4, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x1, %eax
movl %eax, 0xc4(%rsp)
jmp 0x164c85d
jmp 0x164c8ba
movl 0x118(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x118(%rsp)
jmp 0x164c1fa
xorl %eax, %eax
movq %rbp, %rsp
popq %rbp
vzeroupper
retq
nopw (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,250 |
int ncnn::unary_op_inplace<ncnn::UnaryOp_x86_avx512_functor::unary_op_reciprocal>(ncnn::Mat&, ncnn::Option const&)
|
static int unary_op_inplace(Mat& a, const Option& opt)
{
Op op;
int w = a.w;
int h = a.h;
int d = a.d;
int channels = a.c;
int elempack = a.elempack;
int size = w * h * d * elempack;
#pragma omp parallel for num_threads(opt.num_threads)
for (int q = 0; q < channels; q++)
{
float* ptr = a.channel(q);
int i = 0;
#if __SSE2__
#if __AVX__
#if __AVX512F__
for (; i + 15 < size; i += 16)
{
__m512 _p = _mm512_loadu_ps(ptr);
_p = op.func_pack16(_p);
_mm512_storeu_ps(ptr, _p);
ptr += 16;
}
#endif // __AVX512F__
for (; i + 7 < size; i += 8)
{
__m256 _p = _mm256_loadu_ps(ptr);
_p = op.func_pack8(_p);
_mm256_storeu_ps(ptr, _p);
ptr += 8;
}
#endif // __AVX__
for (; i + 3 < size; i += 4)
{
__m128 _p = _mm_load_ps(ptr);
_p = op.func_pack4(_p);
_mm_store_ps(ptr, _p);
ptr += 4;
}
#endif // __SSE2__
for (; i < size; i++)
{
*ptr = op.func(*ptr);
ptr++;
}
}
return 0;
}
|
pushq %rbp
movq %rsp, %rbp
andq $-0x40, %rsp
subq $0x2c0, %rsp # imm = 0x2C0
movq %rdi, 0x140(%rsp)
movq %rsi, 0x138(%rsp)
movq 0x140(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x130(%rsp)
movq 0x140(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0x12c(%rsp)
movq 0x140(%rsp), %rax
movl 0x34(%rax), %eax
movl %eax, 0x128(%rsp)
movq 0x140(%rsp), %rax
movl 0x38(%rax), %eax
movl %eax, 0x124(%rsp)
movq 0x140(%rsp), %rax
movl 0x18(%rax), %eax
movl %eax, 0x120(%rsp)
movl 0x130(%rsp), %eax
imull 0x12c(%rsp), %eax
imull 0x128(%rsp), %eax
imull 0x120(%rsp), %eax
movl %eax, 0x11c(%rsp)
movl $0x0, 0x118(%rsp)
movl 0x118(%rsp), %eax
cmpl 0x124(%rsp), %eax
jge 0x164d060
movq 0x140(%rsp), %rcx
movl 0x118(%rsp), %eax
leaq 0xc8(%rsp), %rdx
movq %rdx, 0x158(%rsp)
movq %rcx, 0x150(%rsp)
movl %eax, 0x14c(%rsp)
movq 0x150(%rsp), %rax
movq %rax, 0x48(%rsp)
movb $0x0, 0x14b(%rsp)
movl 0x2c(%rax), %r9d
movl 0x30(%rax), %r8d
movl 0x34(%rax), %edi
movq (%rax), %rsi
movq 0x40(%rax), %rcx
movslq 0x14c(%rsp), %rdx
imulq %rdx, %rcx
imulq 0x10(%rax), %rcx
addq %rcx, %rsi
movq 0x10(%rax), %rdx
movl 0x18(%rax), %ecx
movq 0x20(%rax), %rax
leaq 0xc8(%rsp), %r10
movq %r10, 0x268(%rsp)
movl %r9d, 0x264(%rsp)
movl %r8d, 0x260(%rsp)
movl %edi, 0x25c(%rsp)
movq %rsi, 0x250(%rsp)
movq %rdx, 0x248(%rsp)
movl %ecx, 0x244(%rsp)
movq %rax, 0x238(%rsp)
movq 0x268(%rsp), %rcx
movq %rcx, 0x40(%rsp)
movq 0x250(%rsp), %rax
movq %rax, (%rcx)
movq $0x0, 0x8(%rcx)
movq 0x248(%rsp), %rax
movq %rax, 0x10(%rcx)
movl 0x244(%rsp), %eax
movl %eax, 0x18(%rcx)
movq 0x238(%rsp), %rax
movq %rax, 0x20(%rcx)
movl $0x3, 0x28(%rcx)
movl 0x264(%rsp), %eax
movl %eax, 0x2c(%rcx)
movl 0x260(%rsp), %eax
movl %eax, 0x30(%rcx)
movl $0x1, 0x34(%rcx)
movl 0x25c(%rsp), %eax
movl %eax, 0x38(%rcx)
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rdx
imulq %rdx, %rax
imulq 0x10(%rcx), %rax
movq %rax, 0x278(%rsp)
movl $0x10, 0x274(%rsp)
movq 0x278(%rsp), %rax
movslq 0x274(%rsp), %rdx
addq %rdx, %rax
subq $0x1, %rax
xorl %edx, %edx
subl 0x274(%rsp), %edx
movslq %edx, %rdx
andq %rdx, %rax
xorl %edx, %edx
divq 0x10(%rcx)
movq %rax, %rdx
movq 0x48(%rsp), %rax
movq %rdx, 0x40(%rcx)
movl 0x28(%rax), %ecx
subl $0x1, %ecx
movl %ecx, 0xf0(%rsp)
cmpl $0x4, 0x28(%rax)
jne 0x164cb4e
movq 0x48(%rsp), %rcx
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rcx
imulq %rcx, %rax
movq %rax, 0x108(%rsp)
movb $0x1, 0x14b(%rsp)
testb $0x1, 0x14b(%rsp)
jne 0x164cc7d
leaq 0xc8(%rsp), %rax
movq %rax, 0x168(%rsp)
movq 0x168(%rsp), %rax
movq %rax, 0x298(%rsp)
movq 0x298(%rsp), %rax
movq %rax, 0x38(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x164cc23
movq 0x38(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x294(%rsp) # imm = 0xFFFFFFFF
movl 0x294(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x290(%rsp)
cmpl $0x1, 0x290(%rsp)
jne 0x164cc23
movq 0x38(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x164cbf4
movq 0x38(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x164cbf2
jmp 0x164cc21
movq 0x38(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x2a0(%rsp)
cmpq $0x0, 0x2a0(%rsp)
je 0x164cc1f
movq 0x2a0(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x164cc21
jmp 0x164cc23
movq 0x38(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x164cc7b
movq %rax, %rdi
callq 0x678a0
jmp 0x164cc7d
leaq 0xc8(%rsp), %rax
movq %rax, 0x160(%rsp)
movq 0x160(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x28(%rsp)
leaq 0xc8(%rsp), %rax
movq %rax, 0x170(%rsp)
movq 0x170(%rsp), %rax
movq %rax, 0x288(%rsp)
movq 0x288(%rsp), %rax
movq %rax, 0x30(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x164cd5c
movq 0x30(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x284(%rsp) # imm = 0xFFFFFFFF
movl 0x284(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x280(%rsp)
cmpl $0x1, 0x280(%rsp)
jne 0x164cd5c
movq 0x30(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x164cd2d
movq 0x30(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x164cd2b
jmp 0x164cd5a
movq 0x30(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x2a8(%rsp)
cmpq $0x0, 0x2a8(%rsp)
je 0x164cd58
movq 0x2a8(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x164cd5a
jmp 0x164cd5c
movq 0x30(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x164cdb4
movq %rax, %rdi
callq 0x678a0
movq 0x28(%rsp), %rax
movq %rax, 0x110(%rsp)
movl $0x0, 0xc4(%rsp)
movl 0xc4(%rsp), %eax
addl $0xf, %eax
cmpl 0x11c(%rsp), %eax
jge 0x164ce86
movq 0x110(%rsp), %rax
movq %rax, 0x178(%rsp)
movq 0x178(%rsp), %rax
vmovups (%rax), %zmm0
vmovaps %zmm0, 0x80(%rsp)
leaq 0x137(%rsp), %rdi
leaq 0x80(%rsp), %rsi
callq 0x16585d0
vmovaps %zmm0, 0x80(%rsp)
movq 0x110(%rsp), %rax
vmovaps 0x80(%rsp), %zmm0
movq %rax, 0x1d0(%rsp)
vmovaps %zmm0, 0x180(%rsp)
vmovaps 0x180(%rsp), %zmm0
movq 0x1d0(%rsp), %rax
vmovups %zmm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x40, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x10, %eax
movl %eax, 0xc4(%rsp)
jmp 0x164cdcc
jmp 0x164ce88
movl 0xc4(%rsp), %eax
addl $0x7, %eax
cmpl 0x11c(%rsp), %eax
jge 0x164cf37
movq 0x110(%rsp), %rax
movq %rax, 0x1d8(%rsp)
movq 0x1d8(%rsp), %rax
vmovups (%rax), %ymm0
vmovaps %ymm0, 0x60(%rsp)
leaq 0x137(%rsp), %rdi
leaq 0x60(%rsp), %rsi
callq 0x1658630
vmovaps %ymm0, 0x60(%rsp)
movq 0x110(%rsp), %rax
vmovaps 0x60(%rsp), %ymm0
movq %rax, 0x210(%rsp)
vmovaps %ymm0, 0x1e0(%rsp)
vmovaps 0x1e0(%rsp), %ymm0
movq 0x210(%rsp), %rax
vmovups %ymm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x20, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x8, %eax
movl %eax, 0xc4(%rsp)
jmp 0x164ce88
jmp 0x164cf39
movl 0xc4(%rsp), %eax
addl $0x3, %eax
cmpl 0x11c(%rsp), %eax
jge 0x164cfeb
movq 0x110(%rsp), %rax
movq %rax, 0x218(%rsp)
movq 0x218(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm0, 0x50(%rsp)
leaq 0x137(%rsp), %rdi
leaq 0x50(%rsp), %rsi
vzeroupper
callq 0x1658680
vmovaps %xmm0, 0x50(%rsp)
movq 0x110(%rsp), %rax
vmovaps 0x50(%rsp), %xmm0
movq %rax, 0x230(%rsp)
vmovaps %xmm0, 0x220(%rsp)
vmovaps 0x220(%rsp), %xmm0
movq 0x230(%rsp), %rax
vmovaps %xmm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x10, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x4, %eax
movl %eax, 0xc4(%rsp)
jmp 0x164cf39
jmp 0x164cfed
movl 0xc4(%rsp), %eax
cmpl 0x11c(%rsp), %eax
jge 0x164d048
movq 0x110(%rsp), %rsi
leaq 0x137(%rsp), %rdi
vzeroupper
callq 0x16586c0
movq 0x110(%rsp), %rax
vmovss %xmm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x4, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x1, %eax
movl %eax, 0xc4(%rsp)
jmp 0x164cfed
jmp 0x164d04a
movl 0x118(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x118(%rsp)
jmp 0x164c98a
xorl %eax, %eax
movq %rbp, %rsp
popq %rbp
vzeroupper
retq
nopw (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,251 |
int ncnn::unary_op_inplace<ncnn::UnaryOp_x86_avx512_functor::unary_op_tanh>(ncnn::Mat&, ncnn::Option const&)
|
static int unary_op_inplace(Mat& a, const Option& opt)
{
Op op;
int w = a.w;
int h = a.h;
int d = a.d;
int channels = a.c;
int elempack = a.elempack;
int size = w * h * d * elempack;
#pragma omp parallel for num_threads(opt.num_threads)
for (int q = 0; q < channels; q++)
{
float* ptr = a.channel(q);
int i = 0;
#if __SSE2__
#if __AVX__
#if __AVX512F__
for (; i + 15 < size; i += 16)
{
__m512 _p = _mm512_loadu_ps(ptr);
_p = op.func_pack16(_p);
_mm512_storeu_ps(ptr, _p);
ptr += 16;
}
#endif // __AVX512F__
for (; i + 7 < size; i += 8)
{
__m256 _p = _mm256_loadu_ps(ptr);
_p = op.func_pack8(_p);
_mm256_storeu_ps(ptr, _p);
ptr += 8;
}
#endif // __AVX__
for (; i + 3 < size; i += 4)
{
__m128 _p = _mm_load_ps(ptr);
_p = op.func_pack4(_p);
_mm_store_ps(ptr, _p);
ptr += 4;
}
#endif // __SSE2__
for (; i < size; i++)
{
*ptr = op.func(*ptr);
ptr++;
}
}
return 0;
}
|
pushq %rbp
movq %rsp, %rbp
andq $-0x40, %rsp
subq $0x2c0, %rsp # imm = 0x2C0
movq %rdi, 0x140(%rsp)
movq %rsi, 0x138(%rsp)
movq 0x140(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x130(%rsp)
movq 0x140(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0x12c(%rsp)
movq 0x140(%rsp), %rax
movl 0x34(%rax), %eax
movl %eax, 0x128(%rsp)
movq 0x140(%rsp), %rax
movl 0x38(%rax), %eax
movl %eax, 0x124(%rsp)
movq 0x140(%rsp), %rax
movl 0x18(%rax), %eax
movl %eax, 0x120(%rsp)
movl 0x130(%rsp), %eax
imull 0x12c(%rsp), %eax
imull 0x128(%rsp), %eax
imull 0x120(%rsp), %eax
movl %eax, 0x11c(%rsp)
movl $0x0, 0x118(%rsp)
movl 0x118(%rsp), %eax
cmpl 0x124(%rsp), %eax
jge 0x164d7f0
movq 0x140(%rsp), %rcx
movl 0x118(%rsp), %eax
leaq 0xc8(%rsp), %rdx
movq %rdx, 0x158(%rsp)
movq %rcx, 0x150(%rsp)
movl %eax, 0x14c(%rsp)
movq 0x150(%rsp), %rax
movq %rax, 0x48(%rsp)
movb $0x0, 0x14b(%rsp)
movl 0x2c(%rax), %r9d
movl 0x30(%rax), %r8d
movl 0x34(%rax), %edi
movq (%rax), %rsi
movq 0x40(%rax), %rcx
movslq 0x14c(%rsp), %rdx
imulq %rdx, %rcx
imulq 0x10(%rax), %rcx
addq %rcx, %rsi
movq 0x10(%rax), %rdx
movl 0x18(%rax), %ecx
movq 0x20(%rax), %rax
leaq 0xc8(%rsp), %r10
movq %r10, 0x268(%rsp)
movl %r9d, 0x264(%rsp)
movl %r8d, 0x260(%rsp)
movl %edi, 0x25c(%rsp)
movq %rsi, 0x250(%rsp)
movq %rdx, 0x248(%rsp)
movl %ecx, 0x244(%rsp)
movq %rax, 0x238(%rsp)
movq 0x268(%rsp), %rcx
movq %rcx, 0x40(%rsp)
movq 0x250(%rsp), %rax
movq %rax, (%rcx)
movq $0x0, 0x8(%rcx)
movq 0x248(%rsp), %rax
movq %rax, 0x10(%rcx)
movl 0x244(%rsp), %eax
movl %eax, 0x18(%rcx)
movq 0x238(%rsp), %rax
movq %rax, 0x20(%rcx)
movl $0x3, 0x28(%rcx)
movl 0x264(%rsp), %eax
movl %eax, 0x2c(%rcx)
movl 0x260(%rsp), %eax
movl %eax, 0x30(%rcx)
movl $0x1, 0x34(%rcx)
movl 0x25c(%rsp), %eax
movl %eax, 0x38(%rcx)
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rdx
imulq %rdx, %rax
imulq 0x10(%rcx), %rax
movq %rax, 0x278(%rsp)
movl $0x10, 0x274(%rsp)
movq 0x278(%rsp), %rax
movslq 0x274(%rsp), %rdx
addq %rdx, %rax
subq $0x1, %rax
xorl %edx, %edx
subl 0x274(%rsp), %edx
movslq %edx, %rdx
andq %rdx, %rax
xorl %edx, %edx
divq 0x10(%rcx)
movq %rax, %rdx
movq 0x48(%rsp), %rax
movq %rdx, 0x40(%rcx)
movl 0x28(%rax), %ecx
subl $0x1, %ecx
movl %ecx, 0xf0(%rsp)
cmpl $0x4, 0x28(%rax)
jne 0x164d2de
movq 0x48(%rsp), %rcx
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rcx
imulq %rcx, %rax
movq %rax, 0x108(%rsp)
movb $0x1, 0x14b(%rsp)
testb $0x1, 0x14b(%rsp)
jne 0x164d40d
leaq 0xc8(%rsp), %rax
movq %rax, 0x168(%rsp)
movq 0x168(%rsp), %rax
movq %rax, 0x298(%rsp)
movq 0x298(%rsp), %rax
movq %rax, 0x38(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x164d3b3
movq 0x38(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x294(%rsp) # imm = 0xFFFFFFFF
movl 0x294(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x290(%rsp)
cmpl $0x1, 0x290(%rsp)
jne 0x164d3b3
movq 0x38(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x164d384
movq 0x38(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x164d382
jmp 0x164d3b1
movq 0x38(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x2a0(%rsp)
cmpq $0x0, 0x2a0(%rsp)
je 0x164d3af
movq 0x2a0(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x164d3b1
jmp 0x164d3b3
movq 0x38(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x164d40b
movq %rax, %rdi
callq 0x678a0
jmp 0x164d40d
leaq 0xc8(%rsp), %rax
movq %rax, 0x160(%rsp)
movq 0x160(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x28(%rsp)
leaq 0xc8(%rsp), %rax
movq %rax, 0x170(%rsp)
movq 0x170(%rsp), %rax
movq %rax, 0x288(%rsp)
movq 0x288(%rsp), %rax
movq %rax, 0x30(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x164d4ec
movq 0x30(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x284(%rsp) # imm = 0xFFFFFFFF
movl 0x284(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x280(%rsp)
cmpl $0x1, 0x280(%rsp)
jne 0x164d4ec
movq 0x30(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x164d4bd
movq 0x30(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x164d4bb
jmp 0x164d4ea
movq 0x30(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x2a8(%rsp)
cmpq $0x0, 0x2a8(%rsp)
je 0x164d4e8
movq 0x2a8(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x164d4ea
jmp 0x164d4ec
movq 0x30(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x164d544
movq %rax, %rdi
callq 0x678a0
movq 0x28(%rsp), %rax
movq %rax, 0x110(%rsp)
movl $0x0, 0xc4(%rsp)
movl 0xc4(%rsp), %eax
addl $0xf, %eax
cmpl 0x11c(%rsp), %eax
jge 0x164d616
movq 0x110(%rsp), %rax
movq %rax, 0x178(%rsp)
movq 0x178(%rsp), %rax
vmovups (%rax), %zmm0
vmovaps %zmm0, 0x80(%rsp)
leaq 0x137(%rsp), %rdi
leaq 0x80(%rsp), %rsi
callq 0x16586e0
vmovaps %zmm0, 0x80(%rsp)
movq 0x110(%rsp), %rax
vmovaps 0x80(%rsp), %zmm0
movq %rax, 0x1d0(%rsp)
vmovaps %zmm0, 0x180(%rsp)
vmovaps 0x180(%rsp), %zmm0
movq 0x1d0(%rsp), %rax
vmovups %zmm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x40, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x10, %eax
movl %eax, 0xc4(%rsp)
jmp 0x164d55c
jmp 0x164d618
movl 0xc4(%rsp), %eax
addl $0x7, %eax
cmpl 0x11c(%rsp), %eax
jge 0x164d6c7
movq 0x110(%rsp), %rax
movq %rax, 0x1d8(%rsp)
movq 0x1d8(%rsp), %rax
vmovups (%rax), %ymm0
vmovaps %ymm0, 0x60(%rsp)
leaq 0x137(%rsp), %rdi
leaq 0x60(%rsp), %rsi
callq 0x1658f00
vmovaps %ymm0, 0x60(%rsp)
movq 0x110(%rsp), %rax
vmovaps 0x60(%rsp), %ymm0
movq %rax, 0x210(%rsp)
vmovaps %ymm0, 0x1e0(%rsp)
vmovaps 0x1e0(%rsp), %ymm0
movq 0x210(%rsp), %rax
vmovups %ymm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x20, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x8, %eax
movl %eax, 0xc4(%rsp)
jmp 0x164d618
jmp 0x164d6c9
movl 0xc4(%rsp), %eax
addl $0x3, %eax
cmpl 0x11c(%rsp), %eax
jge 0x164d77b
movq 0x110(%rsp), %rax
movq %rax, 0x218(%rsp)
movq 0x218(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm0, 0x50(%rsp)
leaq 0x137(%rsp), %rdi
leaq 0x50(%rsp), %rsi
vzeroupper
callq 0x1659bd0
vmovaps %xmm0, 0x50(%rsp)
movq 0x110(%rsp), %rax
vmovaps 0x50(%rsp), %xmm0
movq %rax, 0x230(%rsp)
vmovaps %xmm0, 0x220(%rsp)
vmovaps 0x220(%rsp), %xmm0
movq 0x230(%rsp), %rax
vmovaps %xmm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x10, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x4, %eax
movl %eax, 0xc4(%rsp)
jmp 0x164d6c9
jmp 0x164d77d
movl 0xc4(%rsp), %eax
cmpl 0x11c(%rsp), %eax
jge 0x164d7d8
movq 0x110(%rsp), %rsi
leaq 0x137(%rsp), %rdi
vzeroupper
callq 0x165a600
movq 0x110(%rsp), %rax
vmovss %xmm0, (%rax)
movq 0x110(%rsp), %rax
addq $0x4, %rax
movq %rax, 0x110(%rsp)
movl 0xc4(%rsp), %eax
addl $0x1, %eax
movl %eax, 0xc4(%rsp)
jmp 0x164d77d
jmp 0x164d7da
movl 0x118(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x118(%rsp)
jmp 0x164d11a
xorl %eax, %eax
movq %rbp, %rsp
popq %rbp
vzeroupper
retq
nopw (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,252 |
virtual thunk to ncnn::UnaryOp_x86_avx512::forward_inplace(ncnn::Mat&, ncnn::Option const&) const
|
int UnaryOp_x86_avx512::forward_inplace(Mat& bottom_top_blob, const Option& opt) const
{
using namespace UnaryOp_x86_avx512_functor;
if (op_type == Operation_ABS)
return unary_op_inplace<unary_op_abs>(bottom_top_blob, opt);
if (op_type == Operation_NEG)
return unary_op_inplace<unary_op_neg>(bottom_top_blob, opt);
if (op_type == Operation_FLOOR)
return unary_op_inplace<unary_op_floor>(bottom_top_blob, opt);
if (op_type == Operation_CEIL)
return unary_op_inplace<unary_op_ceil>(bottom_top_blob, opt);
if (op_type == Operation_SQUARE)
return unary_op_inplace<unary_op_square>(bottom_top_blob, opt);
if (op_type == Operation_SQRT)
return unary_op_inplace<unary_op_sqrt>(bottom_top_blob, opt);
if (op_type == Operation_RSQRT)
return unary_op_inplace<unary_op_rsqrt>(bottom_top_blob, opt);
if (op_type == Operation_EXP)
return unary_op_inplace<unary_op_exp>(bottom_top_blob, opt);
if (op_type == Operation_LOG)
return unary_op_inplace<unary_op_log>(bottom_top_blob, opt);
if (op_type == Operation_SIN)
return unary_op_inplace<unary_op_sin>(bottom_top_blob, opt);
if (op_type == Operation_COS)
return unary_op_inplace<unary_op_cos>(bottom_top_blob, opt);
if (op_type == Operation_TAN)
return unary_op_inplace<unary_op_tan>(bottom_top_blob, opt);
if (op_type == Operation_ASIN)
return unary_op_inplace<unary_op_asin>(bottom_top_blob, opt);
if (op_type == Operation_ACOS)
return unary_op_inplace<unary_op_acos>(bottom_top_blob, opt);
if (op_type == Operation_ATAN)
return unary_op_inplace<unary_op_atan>(bottom_top_blob, opt);
if (op_type == Operation_RECIPROCAL)
return unary_op_inplace<unary_op_reciprocal>(bottom_top_blob, opt);
if (op_type == Operation_TANH)
return unary_op_inplace<unary_op_tanh>(bottom_top_blob, opt);
return 0;
}
|
movq %rdi, -0x8(%rsp)
movq %rsi, -0x10(%rsp)
movq %rdx, -0x18(%rsp)
movq -0x8(%rsp), %rdi
movq (%rdi), %rax
movq -0x58(%rax), %rax
addq %rax, %rdi
movq -0x10(%rsp), %rsi
movq -0x18(%rsp), %rdx
jmp 0x1645450
nopl (%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,253 |
ncnn::UnaryOp_x86_avx512_functor::unary_op_abs::func_pack16(float vector[16] const&) const
|
__m512 func_pack16(const __m512& x) const
{
return abs_avx512(x);
}
|
pushq %rbp
movq %rsp, %rbp
andq $-0x40, %rsp
subq $0x200, %rsp # imm = 0x200
movq %rdi, 0x38(%rsp)
movq %rsi, 0x30(%rsp)
movq 0x30(%rsp), %rax
vmovaps (%rax), %zmm0
vmovaps %zmm0, 0x40(%rsp)
vmovaps 0x40(%rsp), %zmm0
vmovaps %zmm0, 0x140(%rsp)
vmovdqa64 0x140(%rsp), %zmm1
movl $0x7fffffff, 0x1ec(%rsp) # imm = 0x7FFFFFFF
vpbroadcastd 0x1ec(%rsp), %zmm0
vmovdqa64 %zmm0, 0x180(%rsp)
vmovdqa64 0x180(%rsp), %zmm0
vmovdqa64 %zmm1, 0x100(%rsp)
vmovdqa64 %zmm0, 0xc0(%rsp)
vmovdqa64 0x100(%rsp), %zmm0
vmovdqa64 0xc0(%rsp), %zmm1
vpandd %zmm1, %zmm0, %zmm0
vmovdqa64 %zmm0, 0x80(%rsp)
vmovdqa64 0x80(%rsp), %zmm0
movq %rbp, %rsp
popq %rbp
retq
nopw %cs:(%rax,%rax)
nopl (%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,254 |
ncnn::UnaryOp_x86_avx512_functor::unary_op_abs::func_pack8(float vector[8] const&) const
|
__m256 func_pack8(const __m256& x) const
{
return abs_avx(x);
}
|
pushq %rbp
movq %rsp, %rbp
andq $-0x20, %rsp
subq $0x100, %rsp # imm = 0x100
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
movq 0x10(%rsp), %rax
vmovaps (%rax), %ymm0
vmovaps %ymm0, 0x20(%rsp)
vpxor %xmm0, %xmm0, %xmm0
vmovaps %ymm0, 0xc0(%rsp)
vmovaps 0xc0(%rsp), %ymm1
vmovaps 0x20(%rsp), %ymm0
vmovaps %ymm1, 0xa0(%rsp)
vmovaps %ymm0, 0x80(%rsp)
vmovaps 0xa0(%rsp), %ymm0
vsubps 0x80(%rsp), %ymm0, %ymm1
vmovaps 0x20(%rsp), %ymm0
vmovaps %ymm1, 0x60(%rsp)
vmovaps %ymm0, 0x40(%rsp)
vmovaps 0x60(%rsp), %ymm0
vmovaps 0x40(%rsp), %ymm1
vmaxps %ymm1, %ymm0, %ymm0
movq %rbp, %rsp
popq %rbp
retq
nop
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,255 |
ncnn::UnaryOp_x86_avx512_functor::unary_op_abs::func_pack4(float vector[4] const&) const
|
__m128 func_pack4(const __m128& x) const
{
return abs_sse(x);
}
|
movq %rdi, -0x70(%rsp)
movq %rsi, -0x78(%rsp)
movq -0x78(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm0, -0x58(%rsp)
movl $0x80000000, -0x3c(%rsp) # imm = 0x80000000
vmovss -0x3c(%rsp), %xmm0
vmovss %xmm0, -0x4(%rsp)
vbroadcastss -0x4(%rsp), %xmm0
vmovaps %xmm0, -0x18(%rsp)
vmovaps -0x18(%rsp), %xmm0
vmovaps %xmm0, -0x68(%rsp)
vmovaps -0x68(%rsp), %xmm1
vmovaps -0x58(%rsp), %xmm0
vmovaps %xmm1, -0x28(%rsp)
vmovaps %xmm0, -0x38(%rsp)
vmovdqa -0x28(%rsp), %xmm0
vpternlogq $0xf, %xmm0, %xmm0, %xmm0
vmovaps -0x38(%rsp), %xmm1
vpand %xmm1, %xmm0, %xmm0
retq
nopw %cs:(%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,256 |
ncnn::UnaryOp_x86_avx512_functor::unary_op_abs::func(float const&) const
|
float func(const float& x) const
{
return (float)fabs(x);
}
|
subq $0x18, %rsp
movq %rdi, 0x10(%rsp)
movq %rsi, 0x8(%rsp)
movq 0x8(%rsp), %rax
vmovss (%rax), %xmm0
callq 0x102e2f0
addq $0x18, %rsp
retq
nopw %cs:(%rax,%rax)
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,257 |
ncnn::UnaryOp_x86_avx512_functor::unary_op_neg::func_pack16(float vector[16] const&) const
|
__m512 func_pack16(const __m512& x) const
{
return _mm512_sub_ps(_mm512_setzero_ps(), x);
}
|
pushq %rbp
movq %rsp, %rbp
andq $-0x40, %rsp
subq $0x140, %rsp # imm = 0x140
movq %rdi, 0x38(%rsp)
movq %rsi, 0x30(%rsp)
vpxor %xmm0, %xmm0, %xmm0
vmovaps %zmm0, 0xc0(%rsp)
vmovaps 0xc0(%rsp), %zmm1
movq 0x30(%rsp), %rax
vmovaps (%rax), %zmm0
vmovaps %zmm1, 0x80(%rsp)
vmovaps %zmm0, 0x40(%rsp)
vmovaps 0x80(%rsp), %zmm0
vsubps 0x40(%rsp), %zmm0, %zmm0
movq %rbp, %rsp
popq %rbp
retq
nopl (%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,258 |
ncnn::UnaryOp_x86_avx512_functor::unary_op_neg::func_pack8(float vector[8] const&) const
|
__m256 func_pack8(const __m256& x) const
{
return _mm256_sub_ps(_mm256_setzero_ps(), x);
}
|
pushq %rbp
movq %rsp, %rbp
andq $-0x20, %rsp
subq $0xa0, %rsp
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
vpxor %xmm0, %xmm0, %xmm0
vmovaps %ymm0, 0x60(%rsp)
vmovaps 0x60(%rsp), %ymm1
movq 0x10(%rsp), %rax
vmovaps (%rax), %ymm0
vmovaps %ymm1, 0x40(%rsp)
vmovaps %ymm0, 0x20(%rsp)
vmovaps 0x40(%rsp), %ymm0
vsubps 0x20(%rsp), %ymm0, %ymm0
movq %rbp, %rsp
popq %rbp
retq
nop
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,259 |
ncnn::UnaryOp_x86_avx512_functor::unary_op_neg::func_pack4(float vector[4] const&) const
|
__m128 func_pack4(const __m128& x) const
{
return _mm_sub_ps(_mm_setzero_ps(), x);
}
|
movq %rdi, -0x40(%rsp)
movq %rsi, -0x48(%rsp)
vpxor %xmm0, %xmm0, %xmm0
vmovaps %xmm0, -0x18(%rsp)
vmovaps -0x18(%rsp), %xmm1
movq -0x48(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm1, -0x28(%rsp)
vmovaps %xmm0, -0x38(%rsp)
vmovaps -0x28(%rsp), %xmm0
vsubps -0x38(%rsp), %xmm0, %xmm0
retq
nopl (%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,260 |
ncnn::UnaryOp_x86_avx512_functor::unary_op_neg::func(float const&) const
|
float func(const float& x) const
{
return -x;
}
|
movq %rdi, -0x8(%rsp)
movq %rsi, -0x10(%rsp)
movq -0x10(%rsp), %rax
vmovss (%rax), %xmm0
vmovd %xmm0, %eax
xorl $0x80000000, %eax # imm = 0x80000000
vmovd %eax, %xmm0
retq
nopw %cs:(%rax,%rax)
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,261 |
ncnn::UnaryOp_x86_avx512_functor::unary_op_floor::func_pack16(float vector[16] const&) const
|
__m512 func_pack16(const __m512& x) const
{
return _mm512_roundscale_ps(x, _MM_FROUND_TO_NEG_INF);
}
|
movq %rdi, -0x8(%rsp)
movq %rsi, -0x10(%rsp)
movq -0x10(%rsp), %rax
vmovaps (%rax), %zmm0
vrndscaleps $0x1, %zmm0, %zmm0
retq
nopl (%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,262 |
ncnn::UnaryOp_x86_avx512_functor::unary_op_floor::func_pack8(float vector[8] const&) const
|
__m256 func_pack8(const __m256& x) const
{
return _mm256_floor_ps(x);
}
|
movq %rdi, -0x8(%rsp)
movq %rsi, -0x10(%rsp)
movq -0x10(%rsp), %rax
vmovaps (%rax), %ymm0
vroundps $0x1, %ymm0, %ymm0
retq
nopw (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,263 |
ncnn::UnaryOp_x86_avx512_functor::unary_op_floor::func_pack4(float vector[4] const&) const
|
__m128 func_pack4(const __m128& x) const
{
#if __SSE4_1__
return _mm_floor_ps(x);
#endif // __SSE4_1__
// Use negative zero as the sign bit mask.
const __m128 magic_negative_zero = _mm_set_ps1(-0.0f);
// The smallest float number that have no fractional part. (2^23)
const __m128 magic_smallest_no_fraction = _mm_set_ps1(8388608.0f);
// absolute = abs(x);
__m128 absolute = _mm_andnot_ps(magic_negative_zero, x);
// negative_mask = magic_negative_zero && x;
__m128 negative_mask = _mm_and_ps(magic_negative_zero, x);
// no_fraction = (magic_smallest_no_fraction < absolute);
__m128 no_fraction = _mm_cmplt_ps(magic_smallest_no_fraction, absolute);
// truncated = static_cast<float>(static_cast<uint32_t>(absolute));
__m128 truncated = _mm_cvtepi32_ps(_mm_cvttps_epi32(absolute));
// truncated_with_sign = (truncated || negative_mask);
__m128 truncated_with_sign = _mm_or_ps(truncated, negative_mask);
// negative_fix = ((x < truncated_with_sign) ? 1.0f : 0.0f);
__m128 negative_fix = _mm_and_ps(
_mm_cmplt_ps(x, truncated_with_sign),
_mm_set_ps1(1.0f));
// fixed_result = truncated_with_sign - negative_fix;
__m128 fixed_result = _mm_sub_ps(truncated_with_sign, negative_fix);
// return ((x && no_fraction) || (!no_fraction && fixed_result));
return _mm_or_ps(
_mm_and_ps(x, no_fraction),
_mm_andnot_ps(no_fraction, fixed_result));
}
|
subq $0x28, %rsp
movq %rdi, 0x20(%rsp)
movq %rsi, 0x18(%rsp)
movq 0x18(%rsp), %rax
vmovaps (%rax), %xmm0
vroundps $0x1, %xmm0, %xmm0
addq $0x28, %rsp
retq
nopw %cs:(%rax,%rax)
nopl (%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,264 |
ncnn::UnaryOp_x86_avx512_functor::unary_op_floor::func(float const&) const
|
float func(const float& x) const
{
return (float)floor(x);
}
|
subq $0x18, %rsp
movq %rdi, 0x10(%rsp)
movq %rsi, 0x8(%rsp)
movq 0x8(%rsp), %rax
vmovss (%rax), %xmm0
callq 0x8e720
addq $0x18, %rsp
retq
nopw %cs:(%rax,%rax)
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,265 |
ncnn::UnaryOp_x86_avx512_functor::unary_op_ceil::func_pack16(float vector[16] const&) const
|
__m512 func_pack16(const __m512& x) const
{
return _mm512_roundscale_ps(x, _MM_FROUND_TO_POS_INF);
}
|
movq %rdi, -0x8(%rsp)
movq %rsi, -0x10(%rsp)
movq -0x10(%rsp), %rax
vmovaps (%rax), %zmm0
vrndscaleps $0x2, %zmm0, %zmm0
retq
nopl (%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,266 |
ncnn::UnaryOp_x86_avx512_functor::unary_op_ceil::func_pack8(float vector[8] const&) const
|
__m256 func_pack8(const __m256& x) const
{
return _mm256_ceil_ps(x);
}
|
movq %rdi, -0x8(%rsp)
movq %rsi, -0x10(%rsp)
movq -0x10(%rsp), %rax
vmovaps (%rax), %ymm0
vroundps $0x2, %ymm0, %ymm0
retq
nopw (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,267 |
ncnn::UnaryOp_x86_avx512_functor::unary_op_ceil::func_pack4(float vector[4] const&) const
|
__m128 func_pack4(const __m128& x) const
{
#if __SSE4_1__
return _mm_ceil_ps(x);
#endif // __SSE4_1__
// Use negative zero as the sign bit mask.
const __m128 magic_negative_zero = _mm_set_ps1(-0.0f);
// The smallest float number that have no fractional part. (2^23)
const __m128 magic_smallest_no_fraction = _mm_set_ps1(8388608.0f);
// absolute = abs(x);
__m128 absolute = _mm_andnot_ps(magic_negative_zero, x);
// negative_mask = magic_negative_zero && x;
__m128 negative_mask = _mm_and_ps(magic_negative_zero, x);
// no_fraction = (magic_smallest_no_fraction < absolute);
__m128 no_fraction = _mm_cmplt_ps(magic_smallest_no_fraction, absolute);
// truncated = static_cast<float>(static_cast<uint32_t>(absolute));
__m128 truncated = _mm_cvtepi32_ps(_mm_cvttps_epi32(absolute));
// truncated_with_sign = (truncated || negative_mask);
__m128 truncated_with_sign = _mm_or_ps(truncated, negative_mask);
// positive_fix = ((x > -0.0f) && (x > truncated_with_sign) ? -1.0f : 0.0f);
__m128 positive_fix = _mm_and_ps(
_mm_and_ps(
_mm_cmpgt_ps(x, magic_negative_zero),
_mm_cmpgt_ps(x, truncated_with_sign)),
_mm_set_ps1(-1.0f));
// fixed_result = truncated_with_sign - positive_fix;
__m128 fixed_result = _mm_sub_ps(truncated_with_sign, positive_fix);
// return ((x && no_fraction) || (!no_fraction && fixed_result));
return _mm_or_ps(
_mm_and_ps(x, no_fraction),
_mm_andnot_ps(no_fraction, fixed_result));
}
|
subq $0x28, %rsp
movq %rdi, 0x20(%rsp)
movq %rsi, 0x18(%rsp)
movq 0x18(%rsp), %rax
vmovaps (%rax), %xmm0
vroundps $0x2, %xmm0, %xmm0
addq $0x28, %rsp
retq
nopw %cs:(%rax,%rax)
nopl (%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,268 |
ncnn::UnaryOp_x86_avx512_functor::unary_op_ceil::func(float const&) const
|
float func(const float& x) const
{
return (float)ceil(x);
}
|
subq $0x18, %rsp
movq %rdi, 0x10(%rsp)
movq %rsi, 0x8(%rsp)
movq 0x8(%rsp), %rax
vmovss (%rax), %xmm0
callq 0x1058980
addq $0x18, %rsp
retq
nopw %cs:(%rax,%rax)
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,269 |
ncnn::UnaryOp_x86_avx512_functor::unary_op_square::func_pack16(float vector[16] const&) const
|
__m512 func_pack16(const __m512& x) const
{
return _mm512_mul_ps(x, x);
}
|
pushq %rbp
movq %rsp, %rbp
andq $-0x40, %rsp
subq $0x100, %rsp # imm = 0x100
movq %rdi, 0x38(%rsp)
movq %rsi, 0x30(%rsp)
movq 0x30(%rsp), %rax
vmovaps (%rax), %zmm1
movq 0x30(%rsp), %rax
vmovaps (%rax), %zmm0
vmovaps %zmm1, 0x80(%rsp)
vmovaps %zmm0, 0x40(%rsp)
vmovaps 0x80(%rsp), %zmm0
vmulps 0x40(%rsp), %zmm0, %zmm0
movq %rbp, %rsp
popq %rbp
retq
nopw %cs:(%rax,%rax)
nop
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,270 |
ncnn::UnaryOp_x86_avx512_functor::unary_op_square::func_pack8(float vector[8] const&) const
|
__m256 func_pack8(const __m256& x) const
{
return _mm256_mul_ps(x, x);
}
|
pushq %rbp
movq %rsp, %rbp
andq $-0x20, %rsp
subq $0x80, %rsp
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
movq 0x10(%rsp), %rax
vmovaps (%rax), %ymm1
movq 0x10(%rsp), %rax
vmovaps (%rax), %ymm0
vmovaps %ymm1, 0x40(%rsp)
vmovaps %ymm0, 0x20(%rsp)
vmovaps 0x40(%rsp), %ymm0
vmulps 0x20(%rsp), %ymm0, %ymm0
movq %rbp, %rsp
popq %rbp
retq
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,271 |
ncnn::UnaryOp_x86_avx512_functor::unary_op_square::func_pack4(float vector[4] const&) const
|
__m128 func_pack4(const __m128& x) const
{
return _mm_mul_ps(x, x);
}
|
movq %rdi, -0x30(%rsp)
movq %rsi, -0x38(%rsp)
movq -0x38(%rsp), %rax
vmovaps (%rax), %xmm1
movq -0x38(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm1, -0x18(%rsp)
vmovaps %xmm0, -0x28(%rsp)
vmovaps -0x18(%rsp), %xmm0
vmulps -0x28(%rsp), %xmm0, %xmm0
retq
nopw %cs:(%rax,%rax)
nop
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,272 |
ncnn::UnaryOp_x86_avx512_functor::unary_op_square::func(float const&) const
|
float func(const float& x) const
{
return x * x;
}
|
movq %rdi, -0x8(%rsp)
movq %rsi, -0x10(%rsp)
movq -0x10(%rsp), %rax
vmovss (%rax), %xmm0
movq -0x10(%rsp), %rax
vmulss (%rax), %xmm0, %xmm0
retq
nopl (%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,273 |
ncnn::UnaryOp_x86_avx512_functor::unary_op_sqrt::func_pack16(float vector[16] const&) const
|
__m512 func_pack16(const __m512& x) const
{
return _mm512_sqrt_ps(x);
}
|
pushq %rbp
movq %rsp, %rbp
andq $-0x40, %rsp
subq $0xc0, %rsp
movq %rdi, 0x38(%rsp)
movq %rsi, 0x30(%rsp)
movq 0x30(%rsp), %rax
vmovaps (%rax), %zmm0
vmovaps %zmm0, 0x40(%rsp)
vmovaps 0x40(%rsp), %zmm2
vrsqrt14ps %zmm2, %zmm1
vmulps %zmm1, %zmm2, %zmm0
vbroadcastss 0x7cefda(%rip), %zmm3 # 0x1e1cdb4
vfmadd213ps %zmm3, %zmm0, %zmm1 # zmm1 = (zmm0 * zmm1) + zmm3
vpbroadcastd 0x7cefc6(%rip), %zmm3 # 0x1e1cdb0
vpandd %zmm3, %zmm2, %zmm3
vbroadcastss 0x7bdfee(%rip), %zmm2 # 0x1e0bde8
vcmpleps %zmm3, %zmm2, %k1
vbroadcastss 0x7bb4a5(%rip), %zmm2 # 0x1e092b0
vmulps %zmm2, %zmm0, %zmm0
vmulps %zmm1, %zmm0, %zmm0 {%k1} {z}
movq %rbp, %rsp
popq %rbp
retq
nopl (%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,274 |
ncnn::UnaryOp_x86_avx512_functor::unary_op_sqrt::func_pack8(float vector[8] const&) const
|
__m256 func_pack8(const __m256& x) const
{
return _mm256_sqrt_ps(x);
}
|
pushq %rbp
movq %rsp, %rbp
andq $-0x20, %rsp
subq $0x60, %rsp
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
movq 0x10(%rsp), %rax
vmovaps (%rax), %ymm0
vmovaps %ymm0, 0x20(%rsp)
vmovaps 0x20(%rsp), %ymm2
vrsqrtps %ymm2, %ymm1
vmulps %ymm1, %ymm2, %ymm0
vbroadcastss 0x7cef58(%rip), %ymm3 # 0x1e1cdb4
vfmadd213ps %ymm3, %ymm0, %ymm1 # ymm1 = (ymm0 * ymm1) + ymm3
vpbroadcastd 0x7cef46(%rip), %ymm3 # 0x1e1cdb0
vpand %ymm3, %ymm2, %ymm3
vbroadcastss 0x7bdf71(%rip), %ymm2 # 0x1e0bde8
vcmpleps %ymm3, %ymm2, %k1
vbroadcastss 0x7bb429(%rip), %ymm2 # 0x1e092b0
vmulps %ymm2, %ymm0, %ymm0
vmulps %ymm1, %ymm0, %ymm0 {%k1} {z}
movq %rbp, %rsp
popq %rbp
retq
nopw %cs:(%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,275 |
ncnn::UnaryOp_x86_avx512_functor::unary_op_sqrt::func_pack4(float vector[4] const&) const
|
__m128 func_pack4(const __m128& x) const
{
return _mm_sqrt_ps(x);
}
|
movq %rdi, -0x20(%rsp)
movq %rsi, -0x28(%rsp)
movq -0x28(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm0, -0x18(%rsp)
vmovaps -0x18(%rsp), %xmm2
vrsqrtps %xmm2, %xmm1
vmulps %xmm1, %xmm2, %xmm0
vbroadcastss 0x7ceee4(%rip), %xmm3 # 0x1e1cdb4
vfmadd213ps %xmm3, %xmm0, %xmm1 # xmm1 = (xmm0 * xmm1) + xmm3
vpbroadcastd 0x7ceed2(%rip), %xmm3 # 0x1e1cdb0
vpand %xmm3, %xmm2, %xmm3
vbroadcastss 0x7bdefd(%rip), %xmm2 # 0x1e0bde8
vcmpleps %xmm3, %xmm2, %k1
vbroadcastss 0x7bb3b5(%rip), %xmm2 # 0x1e092b0
vmulps %xmm2, %xmm0, %xmm0
vmulps %xmm1, %xmm0, %xmm0 {%k1} {z}
retq
nopw %cs:(%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,276 |
ncnn::UnaryOp_x86_avx512_functor::unary_op_sqrt::func(float const&) const
|
float func(const float& x) const
{
return (float)sqrt(x);
}
|
subq $0x18, %rsp
movq %rdi, 0x10(%rsp)
movq %rsi, 0x8(%rsp)
movq 0x8(%rsp), %rax
vmovss (%rax), %xmm0
callq 0x671b0
addq $0x18, %rsp
retq
nopw %cs:(%rax,%rax)
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,277 |
ncnn::UnaryOp_x86_avx512_functor::unary_op_rsqrt::func_pack16(float vector[16] const&) const
|
__m512 func_pack16(const __m512& x) const
{
__m256 _x0 = _mm512_extractf32x8_ps(x, 0);
__m256 _x1 = _mm512_extractf32x8_ps(x, 1);
_x0 = _mm256_rsqrt_ps(_x0);
_x1 = _mm256_rsqrt_ps(_x1);
return _mm512_insertf32x8(_mm512_castps256_ps512(_x0), _x1, 1);
}
|
pushq %rbp
movq %rsp, %rbp
andq $-0x20, %rsp
subq $0xe0, %rsp
movq %rdi, 0x58(%rsp)
movq %rsi, 0x50(%rsp)
movq 0x50(%rsp), %rax
vmovaps (%rax), %zmm0
vmovaps %ymm0, 0x20(%rsp)
movq 0x50(%rsp), %rax
vmovaps (%rax), %zmm0
vextractf64x4 $0x1, %zmm0, (%rsp)
vmovaps 0x20(%rsp), %ymm0
vmovaps %ymm0, 0x80(%rsp)
vrsqrtps 0x80(%rsp), %ymm0
vmovaps %ymm0, 0x20(%rsp)
vmovaps (%rsp), %ymm0
vmovaps %ymm0, 0x60(%rsp)
vrsqrtps 0x60(%rsp), %ymm0
vmovaps %ymm0, (%rsp)
vmovaps 0x20(%rsp), %ymm0
vmovaps %ymm0, 0xa0(%rsp)
vmovaps 0xa0(%rsp), %ymm2
vmovaps (%rsp), %ymm1
vmovaps %ymm2, %ymm0
vinsertf64x4 $0x1, %ymm1, %zmm0, %zmm0
movq %rbp, %rsp
popq %rbp
retq
nop
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,278 |
ncnn::UnaryOp_x86_avx512_functor::unary_op_rsqrt::func_pack8(float vector[8] const&) const
|
__m256 func_pack8(const __m256& x) const
{
return _mm256_rsqrt_ps(x);
}
|
pushq %rbp
movq %rsp, %rbp
andq $-0x20, %rsp
subq $0x60, %rsp
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
movq 0x10(%rsp), %rax
vmovaps (%rax), %ymm0
vmovaps %ymm0, 0x20(%rsp)
vmovaps 0x20(%rsp), %ymm0
vrsqrtps %ymm0, %ymm0
movq %rbp, %rsp
popq %rbp
retq
nopw %cs:(%rax,%rax)
nop
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,279 |
ncnn::UnaryOp_x86_avx512_functor::unary_op_rsqrt::func_pack4(float vector[4] const&) const
|
__m128 func_pack4(const __m128& x) const
{
return _mm_rsqrt_ps(x);
}
|
movq %rdi, -0x20(%rsp)
movq %rsi, -0x28(%rsp)
movq -0x28(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm0, -0x18(%rsp)
vmovaps -0x18(%rsp), %xmm0
vrsqrtps %xmm0, %xmm0
retq
nopw %cs:(%rax,%rax)
nop
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,280 |
ncnn::UnaryOp_x86_avx512_functor::unary_op_rsqrt::func(float const&) const
|
float func(const float& x) const
{
return (float)(1.f / sqrt(x));
}
|
subq $0x18, %rsp
movq %rdi, 0x10(%rsp)
movq %rsi, 0x8(%rsp)
movq 0x8(%rsp), %rax
vmovss (%rax), %xmm0
callq 0x671b0
vmovaps %xmm0, %xmm1
vmovss 0x7baf8c(%rip), %xmm0 # 0x1e09004
vdivss %xmm1, %xmm0, %xmm0
addq $0x18, %rsp
retq
nopw %cs:(%rax,%rax)
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,281 |
ncnn::UnaryOp_x86_avx512_functor::unary_op_exp::func_pack16(float vector[16] const&) const
|
__m512 func_pack16(const __m512& x) const
{
return exp512_ps(x);
}
|
pushq %rbp
movq %rsp, %rbp
andq $-0x40, %rsp
subq $0xf40, %rsp # imm = 0xF40
movq %rdi, 0x38(%rsp)
movq %rsi, 0x30(%rsp)
movq 0x30(%rsp), %rax
vmovaps (%rax), %zmm0
vmovaps %zmm0, 0x200(%rsp)
vpxor %xmm0, %xmm0, %xmm0
vmovaps %zmm0, 0x280(%rsp)
vmovaps 0x280(%rsp), %zmm0
vmovaps %zmm0, 0x1c0(%rsp)
vmovaps 0x7ced5e(%rip), %zmm0 # 0x1e1ce40
vmovaps %zmm0, 0x100(%rsp)
vmovaps 0x200(%rsp), %zmm1
vmovaps 0x7ced84(%rip), %zmm0 # 0x1e1ce80
vmovaps %zmm1, 0x400(%rsp)
vmovaps %zmm0, 0x3c0(%rsp)
vmovaps 0x400(%rsp), %zmm0
vmovaps 0x3c0(%rsp), %zmm1
vminps %zmm1, %zmm0, %zmm0
vmovaps %zmm0, 0x200(%rsp)
vmovaps 0x200(%rsp), %zmm1
vmovaps 0x7ced84(%rip), %zmm0 # 0x1e1cec0
vmovaps %zmm1, 0x480(%rsp)
vmovaps %zmm0, 0x440(%rsp)
vmovaps 0x480(%rsp), %zmm0
vmovaps 0x440(%rsp), %zmm1
vmaxps %zmm1, %zmm0, %zmm0
vmovaps %zmm0, 0x200(%rsp)
vmovaps 0x200(%rsp), %zmm2
vmovaps 0x7ced84(%rip), %zmm1 # 0x1e1cf00
vmovaps 0x7cedba(%rip), %zmm0 # 0x1e1cf40
vmovaps %zmm2, 0x9c0(%rsp)
vmovaps %zmm1, 0x980(%rsp)
vmovaps %zmm0, 0x940(%rsp)
vmovaps 0x9c0(%rsp), %zmm1
vmovaps 0x980(%rsp), %zmm0
vmovaps 0x940(%rsp), %zmm2
vfmadd213ps %zmm2, %zmm1, %zmm0 # zmm0 = (zmm1 * zmm0) + zmm2
vmovaps %zmm0, 0x180(%rsp)
vmovaps 0x180(%rsp), %zmm0
vrndscaleps $0x1, %zmm0, %zmm0
vmovaps %zmm0, 0x1c0(%rsp)
vmovaps 0x1c0(%rsp), %zmm1
vmovaps 0x180(%rsp), %zmm0
vcmpltps %zmm1, %zmm0, %k0
kmovw %k0, 0xfe(%rsp)
vmovaps 0x1c0(%rsp), %zmm1
movw 0xfe(%rsp), %ax
vmovaps 0x100(%rsp), %zmm0
vmovaps %zmm1, 0xac0(%rsp)
movw %ax, 0xabe(%rsp)
vmovaps %zmm1, 0xa40(%rsp)
vmovaps %zmm0, 0xa00(%rsp)
vmovaps 0xa40(%rsp), %zmm1
vmovaps 0xa00(%rsp), %zmm0
kmovw 0xabe(%rsp), %k1
vmovaps %zmm1, 0xb40(%rsp)
vmovaps %zmm0, 0xb00(%rsp)
vmovaps 0xb40(%rsp), %zmm1
vmovaps 0xb00(%rsp), %zmm2
vmovaps 0xac0(%rsp), %zmm0
vsubps %zmm2, %zmm1, %zmm0 {%k1}
vmovaps %zmm0, 0x180(%rsp)
vmovaps 0x180(%rsp), %zmm2
vmovaps 0x7cecec(%rip), %zmm1 # 0x1e1cf80
vmovaps 0x200(%rsp), %zmm0
vmovaps %zmm2, 0xcc0(%rsp)
vmovaps %zmm1, 0xc80(%rsp)
vmovaps %zmm0, 0xc40(%rsp)
vmovaps 0xcc0(%rsp), %zmm1
vmovaps 0xc80(%rsp), %zmm0
vmovaps 0xc40(%rsp), %zmm2
vfnmadd213ps %zmm2, %zmm1, %zmm0 # zmm0 = -(zmm1 * zmm0) + zmm2
vmovaps %zmm0, 0x200(%rsp)
vmovaps 0x180(%rsp), %zmm2
vmovaps 0x7cecd4(%rip), %zmm1 # 0x1e1cfc0
vmovaps 0x200(%rsp), %zmm0
vmovaps %zmm2, 0xc00(%rsp)
vmovaps %zmm1, 0xbc0(%rsp)
vmovaps %zmm0, 0xb80(%rsp)
vmovaps 0xc00(%rsp), %zmm1
vmovaps 0xbc0(%rsp), %zmm0
vmovaps 0xb80(%rsp), %zmm2
vfnmadd213ps %zmm2, %zmm1, %zmm0 # zmm0 = -(zmm1 * zmm0) + zmm2
vmovaps %zmm0, 0x200(%rsp)
vmovaps 0x200(%rsp), %zmm0
vmovaps %zmm0, 0x380(%rsp)
vmovaps %zmm0, 0x340(%rsp)
vmovaps 0x380(%rsp), %zmm0
vmovaps 0x340(%rsp), %zmm1
vmulps %zmm1, %zmm0, %zmm0
vmovaps %zmm0, 0x1c0(%rsp)
vmovaps 0x7cec8e(%rip), %zmm0 # 0x1e1d000
vmovaps %zmm0, 0x80(%rsp)
vmovaps 0x80(%rsp), %zmm2
vmovaps 0x200(%rsp), %zmm1
vmovaps 0x7cecac(%rip), %zmm0 # 0x1e1d040
vmovaps %zmm2, 0x900(%rsp)
vmovaps %zmm1, 0x8c0(%rsp)
vmovaps %zmm0, 0x880(%rsp)
vmovaps 0x900(%rsp), %zmm1
vmovaps 0x8c0(%rsp), %zmm0
vmovaps 0x880(%rsp), %zmm2
vfmadd213ps %zmm2, %zmm1, %zmm0 # zmm0 = (zmm1 * zmm0) + zmm2
vmovaps %zmm0, 0x80(%rsp)
vmovaps 0x80(%rsp), %zmm2
vmovaps 0x200(%rsp), %zmm1
vmovaps 0x7cec94(%rip), %zmm0 # 0x1e1d080
vmovaps %zmm2, 0x840(%rsp)
vmovaps %zmm1, 0x800(%rsp)
vmovaps %zmm0, 0x7c0(%rsp)
vmovaps 0x840(%rsp), %zmm1
vmovaps 0x800(%rsp), %zmm0
vmovaps 0x7c0(%rsp), %zmm2
vfmadd213ps %zmm2, %zmm1, %zmm0 # zmm0 = (zmm1 * zmm0) + zmm2
vmovaps %zmm0, 0x80(%rsp)
vmovaps 0x80(%rsp), %zmm2
vmovaps 0x200(%rsp), %zmm1
vmovaps 0x7cec7c(%rip), %zmm0 # 0x1e1d0c0
vmovaps %zmm2, 0x780(%rsp)
vmovaps %zmm1, 0x740(%rsp)
vmovaps %zmm0, 0x700(%rsp)
vmovaps 0x780(%rsp), %zmm1
vmovaps 0x740(%rsp), %zmm0
vmovaps 0x700(%rsp), %zmm2
vfmadd213ps %zmm2, %zmm1, %zmm0 # zmm0 = (zmm1 * zmm0) + zmm2
vmovaps %zmm0, 0x80(%rsp)
vmovaps 0x80(%rsp), %zmm2
vmovaps 0x200(%rsp), %zmm1
vmovaps 0x7cec64(%rip), %zmm0 # 0x1e1d100
vmovaps %zmm2, 0x6c0(%rsp)
vmovaps %zmm1, 0x680(%rsp)
vmovaps %zmm0, 0x640(%rsp)
vmovaps 0x6c0(%rsp), %zmm1
vmovaps 0x680(%rsp), %zmm0
vmovaps 0x640(%rsp), %zmm2
vfmadd213ps %zmm2, %zmm1, %zmm0 # zmm0 = (zmm1 * zmm0) + zmm2
vmovaps %zmm0, 0x80(%rsp)
vmovaps 0x80(%rsp), %zmm2
vmovaps 0x200(%rsp), %zmm1
vmovaps 0x7cec4c(%rip), %zmm0 # 0x1e1d140
vmovaps %zmm2, 0x600(%rsp)
vmovaps %zmm1, 0x5c0(%rsp)
vmovaps %zmm0, 0x580(%rsp)
vmovaps 0x600(%rsp), %zmm1
vmovaps 0x5c0(%rsp), %zmm0
vmovaps 0x580(%rsp), %zmm2
vfmadd213ps %zmm2, %zmm1, %zmm0 # zmm0 = (zmm1 * zmm0) + zmm2
vmovaps %zmm0, 0x80(%rsp)
vmovaps 0x80(%rsp), %zmm2
vmovaps 0x1c0(%rsp), %zmm1
vmovaps 0x200(%rsp), %zmm0
vmovaps %zmm2, 0x540(%rsp)
vmovaps %zmm1, 0x500(%rsp)
vmovaps %zmm0, 0x4c0(%rsp)
vmovaps 0x540(%rsp), %zmm1
vmovaps 0x500(%rsp), %zmm0
vmovaps 0x4c0(%rsp), %zmm2
vfmadd213ps %zmm2, %zmm1, %zmm0 # zmm0 = (zmm1 * zmm0) + zmm2
vmovaps %zmm0, 0x80(%rsp)
vmovaps 0x80(%rsp), %zmm1
vmovaps 0x100(%rsp), %zmm0
vmovaps %zmm1, 0xd40(%rsp)
vmovaps %zmm0, 0xd00(%rsp)
vmovaps 0xd40(%rsp), %zmm0
vmovaps 0xd00(%rsp), %zmm1
vaddps %zmm1, %zmm0, %zmm0
vmovaps %zmm0, 0x80(%rsp)
vmovaps 0x180(%rsp), %zmm0
vmovaps %zmm0, 0xd80(%rsp)
vmovaps 0xd80(%rsp), %zmm0
vpxor %xmm1, %xmm1, %xmm1
vmovdqa64 %zmm1, 0xec0(%rsp)
vmovdqa64 0xec0(%rsp), %zmm1
vcvttps2dq %zmm0, %zmm0
vmovdqa64 %zmm0, 0x140(%rsp)
vmovdqa64 0x140(%rsp), %zmm1
vmovdqa64 0x7ceb6e(%rip), %zmm0 # 0x1e1d180
vmovdqa64 %zmm1, 0xe00(%rsp)
vmovdqa64 %zmm0, 0xdc0(%rsp)
vmovdqa64 0xe00(%rsp), %zmm0
vmovdqa64 0xdc0(%rsp), %zmm1
vpaddd %zmm1, %zmm0, %zmm0
vmovdqa64 %zmm0, 0x140(%rsp)
vmovdqa64 0x140(%rsp), %zmm0
vmovdqa64 %zmm0, 0xe80(%rsp)
movl $0x17, 0xe7c(%rsp)
vmovdqa64 0xe80(%rsp), %zmm0
movl 0xe7c(%rsp), %eax
vmovd %eax, %xmm1
vpslld %xmm1, %zmm0, %zmm0
vmovdqa64 %zmm0, 0x140(%rsp)
vmovdqa64 0x140(%rsp), %zmm0
vmovdqa64 %zmm0, 0x240(%rsp)
vmovdqa64 0x240(%rsp), %zmm0
vmovaps %zmm0, 0x40(%rsp)
vmovaps 0x80(%rsp), %zmm1
vmovaps 0x40(%rsp), %zmm0
vmovaps %zmm1, 0x300(%rsp)
vmovaps %zmm0, 0x2c0(%rsp)
vmovaps 0x300(%rsp), %zmm0
vmulps 0x2c0(%rsp), %zmm0, %zmm0
vmovaps %zmm0, 0x80(%rsp)
vmovaps 0x80(%rsp), %zmm0
movq %rbp, %rsp
popq %rbp
retq
nopw %cs:(%rax,%rax)
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,282 |
ncnn::UnaryOp_x86_avx512_functor::unary_op_exp::func_pack8(float vector[8] const&) const
|
__m256 func_pack8(const __m256& x) const
{
return exp256_ps(x);
}
|
pushq %rbp
movq %rsp, %rbp
andq $-0x20, %rsp
subq $0xb20, %rsp # imm = 0xB20
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
movq 0x10(%rsp), %rax
vmovaps (%rax), %ymm0
vmovaps %ymm0, 0x100(%rsp)
vpxor %xmm0, %xmm0, %xmm0
vmovaps %ymm0, 0x1a0(%rsp)
vmovaps 0x1a0(%rsp), %ymm0
vmovaps %ymm0, 0xe0(%rsp)
vbroadcastss 0x7ba8c1(%rip), %ymm0 # 0x1e09004
vmovaps %ymm0, 0x80(%rsp)
vmovaps 0x100(%rsp), %ymm0
vmovaps %ymm0, 0x260(%rsp)
vbroadcastss 0x7bd659(%rip), %ymm0 # 0x1e0bdc0
vmovaps %ymm0, 0x240(%rsp)
vmovaps 0x260(%rsp), %ymm0
vmovaps 0x240(%rsp), %ymm1
vminps %ymm1, %ymm0, %ymm0
vmovaps %ymm0, 0x100(%rsp)
vmovaps 0x100(%rsp), %ymm0
vmovaps %ymm0, 0x140(%rsp)
vbroadcastss 0x7bd61a(%rip), %ymm0 # 0x1e0bdc4
vmovaps %ymm0, 0x120(%rsp)
vmovaps 0x140(%rsp), %ymm0
vmovaps 0x120(%rsp), %ymm1
vmaxps %ymm1, %ymm0, %ymm0
vmovaps %ymm0, 0x100(%rsp)
vmovaps 0x100(%rsp), %ymm0
vmovaps %ymm0, 0x500(%rsp)
vbroadcastss 0x7bd5db(%rip), %ymm0 # 0x1e0bdc8
vmovaps %ymm0, 0x4e0(%rsp)
vbroadcastss 0x7ba81d(%rip), %ymm0 # 0x1e0901c
vmovaps %ymm0, 0x4c0(%rsp)
vmovaps 0x500(%rsp), %ymm3
vmovaps 0x4e0(%rsp), %ymm2
vmovaps 0x4c0(%rsp), %ymm1
vmovaps %ymm3, 0x760(%rsp)
vmovaps %ymm2, 0x740(%rsp)
vmovaps %ymm1, 0x720(%rsp)
vmovaps 0x760(%rsp), %ymm2
vmovaps 0x740(%rsp), %ymm1
vmovaps 0x720(%rsp), %ymm3
vfmadd213ps %ymm3, %ymm2, %ymm1 # ymm1 = (ymm2 * ymm1) + ymm3
vmovaps %ymm1, 0xc0(%rsp)
vmovaps 0xc0(%rsp), %ymm1
vroundps $0x1, %ymm1, %ymm1
vmovaps %ymm1, 0xe0(%rsp)
vmovaps 0xe0(%rsp), %ymm2
vmovaps 0xc0(%rsp), %ymm1
vcmpltps %ymm2, %ymm1, %ymm1
vmovdqa %ymm1, 0x60(%rsp)
vmovaps 0x60(%rsp), %ymm2
vmovaps 0x80(%rsp), %ymm1
vmovaps %ymm2, 0x540(%rsp)
vmovaps %ymm1, 0x520(%rsp)
vmovdqa 0x540(%rsp), %ymm1
vmovdqa 0x520(%rsp), %ymm2
vpand %ymm2, %ymm1, %ymm1
vmovdqa %ymm1, 0x60(%rsp)
vmovaps 0xe0(%rsp), %ymm2
vmovaps 0x60(%rsp), %ymm1
vmovaps %ymm2, 0x180(%rsp)
vmovaps %ymm1, 0x160(%rsp)
vmovaps 0x180(%rsp), %ymm1
vmovaps 0x160(%rsp), %ymm2
vsubps %ymm2, %ymm1, %ymm1
vmovaps %ymm1, 0xc0(%rsp)
vmovaps 0xc0(%rsp), %ymm2
vmovaps 0x100(%rsp), %ymm1
vmovaps %ymm2, 0x600(%rsp)
vbroadcastss 0x7bd48f(%rip), %ymm2 # 0x1e0bdcc
vmovaps %ymm2, 0x5e0(%rsp)
vmovaps %ymm1, 0x5c0(%rsp)
vmovaps 0x600(%rsp), %ymm3
vmovaps 0x5e0(%rsp), %ymm2
vmovaps 0x5c0(%rsp), %ymm1
vmovaps %ymm3, 0xa00(%rsp)
vmovaps %ymm2, 0x9e0(%rsp)
vmovaps %ymm1, 0x9c0(%rsp)
vmovaps 0xa00(%rsp), %ymm2
vmovaps 0x9e0(%rsp), %ymm1
vmovaps 0x9c0(%rsp), %ymm3
vfnmadd213ps %ymm3, %ymm2, %ymm1 # ymm1 = -(ymm2 * ymm1) + ymm3
vmovaps %ymm1, 0x100(%rsp)
vmovaps 0xc0(%rsp), %ymm2
vmovaps 0x100(%rsp), %ymm1
vmovaps %ymm2, 0x5a0(%rsp)
vbroadcastss 0x7bd3fe(%rip), %ymm2 # 0x1e0bdd0
vmovaps %ymm2, 0x580(%rsp)
vmovaps %ymm1, 0x560(%rsp)
vmovaps 0x5a0(%rsp), %ymm3
vmovaps 0x580(%rsp), %ymm2
vmovaps 0x560(%rsp), %ymm1
vmovaps %ymm3, 0xa60(%rsp)
vmovaps %ymm2, 0xa40(%rsp)
vmovaps %ymm1, 0xa20(%rsp)
vmovaps 0xa60(%rsp), %ymm2
vmovaps 0xa40(%rsp), %ymm1
vmovaps 0xa20(%rsp), %ymm3
vfnmadd213ps %ymm3, %ymm2, %ymm1 # ymm1 = -(ymm2 * ymm1) + ymm3
vmovaps %ymm1, 0x100(%rsp)
vmovaps 0x100(%rsp), %ymm1
vmovaps %ymm1, 0x220(%rsp)
vmovaps %ymm1, 0x200(%rsp)
vmovaps 0x220(%rsp), %ymm1
vmovaps 0x200(%rsp), %ymm2
vmulps %ymm2, %ymm1, %ymm1
vmovaps %ymm1, 0xe0(%rsp)
vbroadcastss 0x7bd34e(%rip), %ymm1 # 0x1e0bdd4
vmovaps %ymm1, 0x40(%rsp)
vmovaps 0x40(%rsp), %ymm2
vmovaps 0x100(%rsp), %ymm1
vmovaps %ymm2, 0x4a0(%rsp)
vmovaps %ymm1, 0x480(%rsp)
vbroadcastss 0x7bd322(%rip), %ymm1 # 0x1e0bdd8
vmovaps %ymm1, 0x460(%rsp)
vmovaps 0x4a0(%rsp), %ymm3
vmovaps 0x480(%rsp), %ymm2
vmovaps 0x460(%rsp), %ymm1
vmovaps %ymm3, 0x7c0(%rsp)
vmovaps %ymm2, 0x7a0(%rsp)
vmovaps %ymm1, 0x780(%rsp)
vmovaps 0x7c0(%rsp), %ymm2
vmovaps 0x7a0(%rsp), %ymm1
vmovaps 0x780(%rsp), %ymm3
vfmadd213ps %ymm3, %ymm2, %ymm1 # ymm1 = (ymm2 * ymm1) + ymm3
vmovaps %ymm1, 0x40(%rsp)
vmovaps 0x40(%rsp), %ymm2
vmovaps 0x100(%rsp), %ymm1
vmovaps %ymm2, 0x440(%rsp)
vmovaps %ymm1, 0x420(%rsp)
vbroadcastss 0x7bd297(%rip), %ymm1 # 0x1e0bddc
vmovaps %ymm1, 0x400(%rsp)
vmovaps 0x440(%rsp), %ymm3
vmovaps 0x420(%rsp), %ymm2
vmovaps 0x400(%rsp), %ymm1
vmovaps %ymm3, 0x820(%rsp)
vmovaps %ymm2, 0x800(%rsp)
vmovaps %ymm1, 0x7e0(%rsp)
vmovaps 0x820(%rsp), %ymm2
vmovaps 0x800(%rsp), %ymm1
vmovaps 0x7e0(%rsp), %ymm3
vfmadd213ps %ymm3, %ymm2, %ymm1 # ymm1 = (ymm2 * ymm1) + ymm3
vmovaps %ymm1, 0x40(%rsp)
vmovaps 0x40(%rsp), %ymm2
vmovaps 0x100(%rsp), %ymm1
vmovaps %ymm2, 0x3e0(%rsp)
vmovaps %ymm1, 0x3c0(%rsp)
vbroadcastss 0x7bd20c(%rip), %ymm1 # 0x1e0bde0
vmovaps %ymm1, 0x3a0(%rsp)
vmovaps 0x3e0(%rsp), %ymm3
vmovaps 0x3c0(%rsp), %ymm2
vmovaps 0x3a0(%rsp), %ymm1
vmovaps %ymm3, 0x880(%rsp)
vmovaps %ymm2, 0x860(%rsp)
vmovaps %ymm1, 0x840(%rsp)
vmovaps 0x880(%rsp), %ymm2
vmovaps 0x860(%rsp), %ymm1
vmovaps 0x840(%rsp), %ymm3
vfmadd213ps %ymm3, %ymm2, %ymm1 # ymm1 = (ymm2 * ymm1) + ymm3
vmovaps %ymm1, 0x40(%rsp)
vmovaps 0x40(%rsp), %ymm2
vmovaps 0x100(%rsp), %ymm1
vmovaps %ymm2, 0x380(%rsp)
vmovaps %ymm1, 0x360(%rsp)
vbroadcastss 0x7bd181(%rip), %ymm1 # 0x1e0bde4
vmovaps %ymm1, 0x340(%rsp)
vmovaps 0x380(%rsp), %ymm3
vmovaps 0x360(%rsp), %ymm2
vmovaps 0x340(%rsp), %ymm1
vmovaps %ymm3, 0x8e0(%rsp)
vmovaps %ymm2, 0x8c0(%rsp)
vmovaps %ymm1, 0x8a0(%rsp)
vmovaps 0x8e0(%rsp), %ymm2
vmovaps 0x8c0(%rsp), %ymm1
vmovaps 0x8a0(%rsp), %ymm3
vfmadd213ps %ymm3, %ymm2, %ymm1 # ymm1 = (ymm2 * ymm1) + ymm3
vmovaps %ymm1, 0x40(%rsp)
vmovaps 0x40(%rsp), %ymm2
vmovaps 0x100(%rsp), %ymm1
vmovaps %ymm2, 0x320(%rsp)
vmovaps %ymm1, 0x300(%rsp)
vmovaps %ymm0, 0x2e0(%rsp)
vmovaps 0x320(%rsp), %ymm2
vmovaps 0x300(%rsp), %ymm1
vmovaps 0x2e0(%rsp), %ymm0
vmovaps %ymm2, 0x940(%rsp)
vmovaps %ymm1, 0x920(%rsp)
vmovaps %ymm0, 0x900(%rsp)
vmovaps 0x940(%rsp), %ymm1
vmovaps 0x920(%rsp), %ymm0
vmovaps 0x900(%rsp), %ymm2
vfmadd213ps %ymm2, %ymm1, %ymm0 # ymm0 = (ymm1 * ymm0) + ymm2
vmovaps %ymm0, 0x40(%rsp)
vmovaps 0x40(%rsp), %ymm2
vmovaps 0xe0(%rsp), %ymm1
vmovaps 0x100(%rsp), %ymm0
vmovaps %ymm2, 0x2c0(%rsp)
vmovaps %ymm1, 0x2a0(%rsp)
vmovaps %ymm0, 0x280(%rsp)
vmovaps 0x2c0(%rsp), %ymm2
vmovaps 0x2a0(%rsp), %ymm1
vmovaps 0x280(%rsp), %ymm0
vmovaps %ymm2, 0x9a0(%rsp)
vmovaps %ymm1, 0x980(%rsp)
vmovaps %ymm0, 0x960(%rsp)
vmovaps 0x9a0(%rsp), %ymm1
vmovaps 0x980(%rsp), %ymm0
vmovaps 0x960(%rsp), %ymm2
vfmadd213ps %ymm2, %ymm1, %ymm0 # ymm0 = (ymm1 * ymm0) + ymm2
vmovaps %ymm0, 0x40(%rsp)
vmovaps 0x40(%rsp), %ymm1
vmovaps 0x80(%rsp), %ymm0
vmovaps %ymm1, 0x640(%rsp)
vmovaps %ymm0, 0x620(%rsp)
vmovaps 0x640(%rsp), %ymm0
vmovaps 0x620(%rsp), %ymm1
vaddps %ymm1, %ymm0, %ymm0
vmovaps %ymm0, 0x40(%rsp)
vmovaps 0xc0(%rsp), %ymm0
vmovaps %ymm0, 0x660(%rsp)
vcvttps2dq 0x660(%rsp), %ymm0
vmovdqa %ymm0, 0xa0(%rsp)
vmovdqa 0xa0(%rsp), %ymm0
vmovdqa %ymm0, 0x6a0(%rsp)
vpbroadcastq 0x7bcfbf(%rip), %ymm0 # 0x1e0be18
vmovdqa %ymm0, 0x680(%rsp)
vmovdqa 0x6a0(%rsp), %ymm1
vmovdqa 0x680(%rsp), %ymm0
vmovdqa %ymm1, 0xaa0(%rsp)
vmovdqa %ymm0, 0xa80(%rsp)
vmovdqa 0xaa0(%rsp), %ymm0
vmovdqa 0xa80(%rsp), %ymm1
vpaddd %ymm1, %ymm0, %ymm0
vmovdqa %ymm0, 0xa0(%rsp)
vmovdqa 0xa0(%rsp), %ymm0
vmovdqa %ymm0, 0x6e0(%rsp)
movl $0x17, 0x6dc(%rsp)
vmovdqa 0x6e0(%rsp), %ymm0
movl 0x6dc(%rsp), %eax
vmovdqa %ymm0, 0xae0(%rsp)
movl %eax, 0xadc(%rsp)
vmovdqa 0xae0(%rsp), %ymm0
movl 0xadc(%rsp), %eax
vmovd %eax, %xmm1
vpslld %xmm1, %ymm0, %ymm0
vmovdqa %ymm0, 0xa0(%rsp)
vmovdqa 0xa0(%rsp), %ymm0
vmovdqa %ymm0, 0x700(%rsp)
vmovdqa 0x700(%rsp), %ymm0
vmovaps %ymm0, 0x20(%rsp)
vmovaps 0x40(%rsp), %ymm1
vmovaps 0x20(%rsp), %ymm0
vmovaps %ymm1, 0x1e0(%rsp)
vmovaps %ymm0, 0x1c0(%rsp)
vmovaps 0x1e0(%rsp), %ymm0
vmulps 0x1c0(%rsp), %ymm0, %ymm0
vmovaps %ymm0, 0x40(%rsp)
vmovaps 0x40(%rsp), %ymm0
movq %rbp, %rsp
popq %rbp
retq
nopw %cs:(%rax,%rax)
nop
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,283 |
ncnn::UnaryOp_x86_avx512_functor::unary_op_exp::func_pack4(float vector[4] const&) const
|
__m128 func_pack4(const __m128& x) const
{
return exp_ps(x);
}
|
subq $0x4e8, %rsp # imm = 0x4E8
movq %rdi, -0x78(%rsp)
movq %rsi, -0x80(%rsp)
movq -0x80(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm0, (%rsp)
vpxor %xmm0, %xmm0, %xmm0
vmovaps %xmm0, 0x30(%rsp)
vmovaps 0x30(%rsp), %xmm0
vmovaps %xmm0, -0x10(%rsp)
vbroadcastss 0x7ba056(%rip), %xmm0 # 0x1e09004
vmovaps %xmm0, -0x40(%rsp)
vmovaps (%rsp), %xmm0
vmovaps %xmm0, 0xb0(%rsp)
vbroadcastss 0x7bcdf5(%rip), %xmm0 # 0x1e0bdc0
vmovaps %xmm0, 0xa0(%rsp)
vmovaps 0xb0(%rsp), %xmm0
vmovaps 0xa0(%rsp), %xmm1
vminps %xmm1, %xmm0, %xmm0
vmovaps %xmm0, (%rsp)
vmovaps (%rsp), %xmm0
vmovaps %xmm0, 0xd0(%rsp)
vbroadcastss 0x7bcdbe(%rip), %xmm0 # 0x1e0bdc4
vmovaps %xmm0, 0xc0(%rsp)
vmovaps 0xd0(%rsp), %xmm0
vmovaps 0xc0(%rsp), %xmm1
vmaxps %xmm1, %xmm0, %xmm0
vmovaps %xmm0, (%rsp)
vmovaps (%rsp), %xmm0
vmovaps %xmm0, 0x90(%rsp)
vbroadcastss 0x7bcd87(%rip), %xmm0 # 0x1e0bdc8
vmovaps %xmm0, 0x80(%rsp)
vmovaps 0x90(%rsp), %xmm0
vmovaps 0x80(%rsp), %xmm1
vmulps %xmm1, %xmm0, %xmm0
vmovaps %xmm0, -0x20(%rsp)
vmovaps -0x20(%rsp), %xmm0
vmovaps %xmm0, 0x110(%rsp)
vbroadcastss 0x7b9f9e(%rip), %xmm0 # 0x1e0901c
vmovaps %xmm0, 0x100(%rsp)
vmovaps 0x110(%rsp), %xmm1
vmovaps 0x100(%rsp), %xmm2
vaddps %xmm2, %xmm1, %xmm1
vmovaps %xmm1, -0x20(%rsp)
vmovaps -0x20(%rsp), %xmm1
vmovaps %xmm1, 0x130(%rsp)
vcvttps2dq 0x130(%rsp), %xmm1
vmovdqa %xmm1, -0x30(%rsp)
vmovdqa -0x30(%rsp), %xmm1
vmovdqa %xmm1, 0x140(%rsp)
vcvtdq2ps 0x140(%rsp), %xmm1
vmovaps %xmm1, -0x10(%rsp)
vmovaps -0x10(%rsp), %xmm2
vmovaps -0x20(%rsp), %xmm1
vmovaps %xmm2, 0x160(%rsp)
vmovaps %xmm1, 0x150(%rsp)
vmovaps 0x150(%rsp), %xmm1
vmovaps 0x160(%rsp), %xmm2
vcmpltps %xmm2, %xmm1, %xmm1
vmovaps %xmm1, -0x50(%rsp)
vmovaps -0x50(%rsp), %xmm2
vmovaps -0x40(%rsp), %xmm1
vmovaps %xmm2, 0x180(%rsp)
vmovaps %xmm1, 0x170(%rsp)
vmovdqa 0x180(%rsp), %xmm1
vmovdqa 0x170(%rsp), %xmm2
vpand %xmm2, %xmm1, %xmm1
vmovdqa %xmm1, -0x50(%rsp)
vmovaps -0x10(%rsp), %xmm2
vmovaps -0x50(%rsp), %xmm1
vmovaps %xmm2, 0x20(%rsp)
vmovaps %xmm1, 0x10(%rsp)
vmovaps 0x20(%rsp), %xmm1
vmovaps 0x10(%rsp), %xmm2
vsubps %xmm2, %xmm1, %xmm1
vmovaps %xmm1, -0x20(%rsp)
vmovaps -0x20(%rsp), %xmm2
vmovaps (%rsp), %xmm1
vmovaps %xmm2, 0x1e0(%rsp)
vbroadcastss 0x7bcc2d(%rip), %xmm2 # 0x1e0bdcc
vmovaps %xmm2, 0x1d0(%rsp)
vmovaps %xmm1, 0x1c0(%rsp)
vmovaps 0x1e0(%rsp), %xmm3
vmovaps 0x1d0(%rsp), %xmm2
vmovaps 0x1c0(%rsp), %xmm1
vmovaps %xmm3, 0x380(%rsp)
vmovaps %xmm2, 0x370(%rsp)
vmovaps %xmm1, 0x360(%rsp)
vmovaps 0x380(%rsp), %xmm2
vmovaps 0x370(%rsp), %xmm1
vmovaps 0x360(%rsp), %xmm3
vfnmadd213ps %xmm3, %xmm2, %xmm1 # xmm1 = -(xmm2 * xmm1) + xmm3
vmovaps %xmm1, (%rsp)
vmovaps -0x20(%rsp), %xmm2
vmovaps (%rsp), %xmm1
vmovaps %xmm2, 0x1b0(%rsp)
vbroadcastss 0x7bcba7(%rip), %xmm2 # 0x1e0bdd0
vmovaps %xmm2, 0x1a0(%rsp)
vmovaps %xmm1, 0x190(%rsp)
vmovaps 0x1b0(%rsp), %xmm3
vmovaps 0x1a0(%rsp), %xmm2
vmovaps 0x190(%rsp), %xmm1
vmovaps %xmm3, 0x3b0(%rsp)
vmovaps %xmm2, 0x3a0(%rsp)
vmovaps %xmm1, 0x390(%rsp)
vmovaps 0x3b0(%rsp), %xmm2
vmovaps 0x3a0(%rsp), %xmm1
vmovaps 0x390(%rsp), %xmm3
vfnmadd213ps %xmm3, %xmm2, %xmm1 # xmm1 = -(xmm2 * xmm1) + xmm3
vmovaps %xmm1, (%rsp)
vmovaps (%rsp), %xmm1
vmovaps %xmm1, 0x70(%rsp)
vmovaps %xmm1, 0x60(%rsp)
vmovaps 0x70(%rsp), %xmm1
vmovaps 0x60(%rsp), %xmm2
vmulps %xmm2, %xmm1, %xmm1
vmovaps %xmm1, -0x10(%rsp)
vbroadcastss 0x7bcb0e(%rip), %xmm1 # 0x1e0bdd4
vmovaps %xmm1, -0x60(%rsp)
vmovaps -0x60(%rsp), %xmm2
vmovaps (%rsp), %xmm1
vmovaps %xmm2, 0x300(%rsp)
vmovaps %xmm1, 0x2f0(%rsp)
vbroadcastss 0x7bcae6(%rip), %xmm1 # 0x1e0bdd8
vmovaps %xmm1, 0x2e0(%rsp)
vmovaps 0x300(%rsp), %xmm3
vmovaps 0x2f0(%rsp), %xmm2
vmovaps 0x2e0(%rsp), %xmm1
vmovaps %xmm3, 0x3e0(%rsp)
vmovaps %xmm2, 0x3d0(%rsp)
vmovaps %xmm1, 0x3c0(%rsp)
vmovaps 0x3e0(%rsp), %xmm2
vmovaps 0x3d0(%rsp), %xmm1
vmovaps 0x3c0(%rsp), %xmm3
vfmadd213ps %xmm3, %xmm2, %xmm1 # xmm1 = (xmm2 * xmm1) + xmm3
vmovaps %xmm1, -0x60(%rsp)
vmovaps -0x60(%rsp), %xmm2
vmovaps (%rsp), %xmm1
vmovaps %xmm2, 0x2d0(%rsp)
vmovaps %xmm1, 0x2c0(%rsp)
vbroadcastss 0x7bca5f(%rip), %xmm1 # 0x1e0bddc
vmovaps %xmm1, 0x2b0(%rsp)
vmovaps 0x2d0(%rsp), %xmm3
vmovaps 0x2c0(%rsp), %xmm2
vmovaps 0x2b0(%rsp), %xmm1
vmovaps %xmm3, 0x410(%rsp)
vmovaps %xmm2, 0x400(%rsp)
vmovaps %xmm1, 0x3f0(%rsp)
vmovaps 0x410(%rsp), %xmm2
vmovaps 0x400(%rsp), %xmm1
vmovaps 0x3f0(%rsp), %xmm3
vfmadd213ps %xmm3, %xmm2, %xmm1 # xmm1 = (xmm2 * xmm1) + xmm3
vmovaps %xmm1, -0x60(%rsp)
vmovaps -0x60(%rsp), %xmm2
vmovaps (%rsp), %xmm1
vmovaps %xmm2, 0x2a0(%rsp)
vmovaps %xmm1, 0x290(%rsp)
vbroadcastss 0x7bc9d8(%rip), %xmm1 # 0x1e0bde0
vmovaps %xmm1, 0x280(%rsp)
vmovaps 0x2a0(%rsp), %xmm3
vmovaps 0x290(%rsp), %xmm2
vmovaps 0x280(%rsp), %xmm1
vmovaps %xmm3, 0x440(%rsp)
vmovaps %xmm2, 0x430(%rsp)
vmovaps %xmm1, 0x420(%rsp)
vmovaps 0x440(%rsp), %xmm2
vmovaps 0x430(%rsp), %xmm1
vmovaps 0x420(%rsp), %xmm3
vfmadd213ps %xmm3, %xmm2, %xmm1 # xmm1 = (xmm2 * xmm1) + xmm3
vmovaps %xmm1, -0x60(%rsp)
vmovaps -0x60(%rsp), %xmm2
vmovaps (%rsp), %xmm1
vmovaps %xmm2, 0x270(%rsp)
vmovaps %xmm1, 0x260(%rsp)
vbroadcastss 0x7bc951(%rip), %xmm1 # 0x1e0bde4
vmovaps %xmm1, 0x250(%rsp)
vmovaps 0x270(%rsp), %xmm3
vmovaps 0x260(%rsp), %xmm2
vmovaps 0x250(%rsp), %xmm1
vmovaps %xmm3, 0x470(%rsp)
vmovaps %xmm2, 0x460(%rsp)
vmovaps %xmm1, 0x450(%rsp)
vmovaps 0x470(%rsp), %xmm2
vmovaps 0x460(%rsp), %xmm1
vmovaps 0x450(%rsp), %xmm3
vfmadd213ps %xmm3, %xmm2, %xmm1 # xmm1 = (xmm2 * xmm1) + xmm3
vmovaps %xmm1, -0x60(%rsp)
vmovaps -0x60(%rsp), %xmm2
vmovaps (%rsp), %xmm1
vmovaps %xmm2, 0x240(%rsp)
vmovaps %xmm1, 0x230(%rsp)
vmovaps %xmm0, 0x220(%rsp)
vmovaps 0x240(%rsp), %xmm2
vmovaps 0x230(%rsp), %xmm1
vmovaps 0x220(%rsp), %xmm0
vmovaps %xmm2, 0x4a0(%rsp)
vmovaps %xmm1, 0x490(%rsp)
vmovaps %xmm0, 0x480(%rsp)
vmovaps 0x4a0(%rsp), %xmm1
vmovaps 0x490(%rsp), %xmm0
vmovaps 0x480(%rsp), %xmm2
vfmadd213ps %xmm2, %xmm1, %xmm0 # xmm0 = (xmm1 * xmm0) + xmm2
vmovaps %xmm0, -0x60(%rsp)
vmovaps -0x60(%rsp), %xmm2
vmovaps -0x10(%rsp), %xmm1
vmovaps (%rsp), %xmm0
vmovaps %xmm2, 0x210(%rsp)
vmovaps %xmm1, 0x200(%rsp)
vmovaps %xmm0, 0x1f0(%rsp)
vmovaps 0x210(%rsp), %xmm2
vmovaps 0x200(%rsp), %xmm1
vmovaps 0x1f0(%rsp), %xmm0
vmovaps %xmm2, 0x4d0(%rsp)
vmovaps %xmm1, 0x4c0(%rsp)
vmovaps %xmm0, 0x4b0(%rsp)
vmovaps 0x4d0(%rsp), %xmm1
vmovaps 0x4c0(%rsp), %xmm0
vmovaps 0x4b0(%rsp), %xmm2
vfmadd213ps %xmm2, %xmm1, %xmm0 # xmm0 = (xmm1 * xmm0) + xmm2
vmovaps %xmm0, -0x60(%rsp)
vmovaps -0x60(%rsp), %xmm1
vmovaps -0x40(%rsp), %xmm0
vmovaps %xmm1, 0xf0(%rsp)
vmovaps %xmm0, 0xe0(%rsp)
vmovaps 0xf0(%rsp), %xmm0
vmovaps 0xe0(%rsp), %xmm1
vaddps %xmm1, %xmm0, %xmm0
vmovaps %xmm0, -0x60(%rsp)
vmovaps -0x20(%rsp), %xmm0
vmovaps %xmm0, 0x120(%rsp)
vcvttps2dq 0x120(%rsp), %xmm0
vmovdqa %xmm0, -0x30(%rsp)
vmovdqa -0x30(%rsp), %xmm0
vmovdqa %xmm0, 0x320(%rsp)
vpbroadcastq 0x7bc7a6(%rip), %xmm0 # 0x1e0be18
vmovdqa %xmm0, 0x310(%rsp)
vmovdqa 0x320(%rsp), %xmm0
vmovdqa 0x310(%rsp), %xmm1
vpaddd %xmm1, %xmm0, %xmm0
vmovdqa %xmm0, -0x30(%rsp)
vmovdqa -0x30(%rsp), %xmm0
vmovdqa %xmm0, 0x340(%rsp)
movl $0x17, 0x33c(%rsp)
vmovdqa 0x340(%rsp), %xmm0
movl 0x33c(%rsp), %eax
vmovd %eax, %xmm1
vpslld %xmm1, %xmm0, %xmm0
vmovdqa %xmm0, -0x30(%rsp)
vmovdqa -0x30(%rsp), %xmm0
vmovdqa %xmm0, 0x350(%rsp)
vmovdqa 0x350(%rsp), %xmm0
vmovaps %xmm0, -0x70(%rsp)
vmovaps -0x60(%rsp), %xmm1
vmovaps -0x70(%rsp), %xmm0
vmovaps %xmm1, 0x50(%rsp)
vmovaps %xmm0, 0x40(%rsp)
vmovaps 0x50(%rsp), %xmm0
vmulps 0x40(%rsp), %xmm0, %xmm0
vmovaps %xmm0, -0x60(%rsp)
vmovaps -0x60(%rsp), %xmm0
addq $0x4e8, %rsp # imm = 0x4E8
retq
nopw %cs:(%rax,%rax)
nop
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,284 |
ncnn::UnaryOp_x86_avx512_functor::unary_op_exp::func(float const&) const
|
float func(const float& x) const
{
return (float)exp(x);
}
|
subq $0x18, %rsp
movq %rdi, 0x10(%rsp)
movq %rsi, 0x8(%rsp)
movq 0x8(%rsp), %rax
vmovss (%rax), %xmm0
callq 0x100cf0
addq $0x18, %rsp
retq
nopw %cs:(%rax,%rax)
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,285 |
ncnn::UnaryOp_x86_avx512_functor::unary_op_log::func_pack16(float vector[16] const&) const
|
__m512 func_pack16(const __m512& x) const
{
return log512_ps(x);
}
|
pushq %rbp
movq %rsp, %rbp
andq $-0x40, %rsp
subq $0x1500, %rsp # imm = 0x1500
movq %rdi, 0x38(%rsp)
movq %rsi, 0x30(%rsp)
movq 0x30(%rsp), %rax
vmovaps (%rax), %zmm0
vmovaps %zmm0, 0x240(%rsp)
vmovaps 0x7cd6aa(%rip), %zmm0 # 0x1e1ce40
vmovaps %zmm0, 0x1c0(%rsp)
vmovaps 0x240(%rsp), %zmm0
vpxor %xmm1, %xmm1, %xmm1
vmovaps %zmm1, 0x380(%rsp)
vmovaps 0x380(%rsp), %zmm1
vcmpleps %zmm1, %zmm0, %k0
kmovw %k0, 0x1be(%rsp)
vmovaps 0x240(%rsp), %zmm1
vmovaps 0x7cdca4(%rip), %zmm0 # 0x1e1d480
vmovaps %zmm1, 0x580(%rsp)
vmovaps %zmm0, 0x540(%rsp)
vmovaps 0x580(%rsp), %zmm0
vmovaps 0x540(%rsp), %zmm1
vmaxps %zmm1, %zmm0, %zmm0
vmovaps %zmm0, 0x240(%rsp)
vmovaps 0x240(%rsp), %zmm0
vmovaps %zmm0, 0x2c0(%rsp)
vmovaps 0x2c0(%rsp), %zmm0
vmovaps %zmm0, 0x10c0(%rsp)
movl $0x17, 0x10bc(%rsp)
vmovdqa64 0x10c0(%rsp), %zmm0
vmovd 0x10bc(%rsp), %xmm1
vpsrld %xmm1, %zmm0, %zmm0
vmovdqa64 %zmm0, 0x200(%rsp)
vmovaps 0x240(%rsp), %zmm1
vmovaps 0x7cdc5a(%rip), %zmm0 # 0x1e1d4c0
vmovaps %zmm1, 0x1140(%rsp)
vmovaps %zmm0, 0x1100(%rsp)
vmovdqa64 0x1140(%rsp), %zmm0
vmovdqa64 0x1100(%rsp), %zmm1
vpandd %zmm1, %zmm0, %zmm0
vmovdqa64 %zmm0, 0x240(%rsp)
vmovaps 0x240(%rsp), %zmm1
vmovaps 0x7cd69a(%rip), %zmm0 # 0x1e1cf40
vmovaps %zmm1, 0x1240(%rsp)
vmovaps %zmm0, 0x1200(%rsp)
vmovdqa64 0x1240(%rsp), %zmm0
vmovdqa64 0x1200(%rsp), %zmm1
vpord %zmm1, %zmm0, %zmm0
vmovdqa64 %zmm0, 0x240(%rsp)
vmovdqa64 0x200(%rsp), %zmm1
vmovdqa64 0x7cd89a(%rip), %zmm0 # 0x1e1d180
vmovdqa64 %zmm1, 0x12c0(%rsp)
vmovdqa64 %zmm0, 0x1280(%rsp)
vmovdqa64 0x12c0(%rsp), %zmm0
vmovdqa64 0x1280(%rsp), %zmm1
vpsubd %zmm1, %zmm0, %zmm0
vmovdqa64 %zmm0, 0x200(%rsp)
vmovdqa64 0x200(%rsp), %zmm0
vmovdqa64 %zmm0, 0x1300(%rsp)
vcvtdq2ps 0x1300(%rsp), %zmm0
vmovaps %zmm0, 0x140(%rsp)
vmovaps 0x140(%rsp), %zmm1
vmovaps 0x1c0(%rsp), %zmm0
vmovaps %zmm1, 0x1040(%rsp)
vmovaps %zmm0, 0x1000(%rsp)
vmovaps 0x1040(%rsp), %zmm0
vmovaps 0x1000(%rsp), %zmm1
vaddps %zmm1, %zmm0, %zmm0
vmovaps %zmm0, 0x140(%rsp)
vmovaps 0x240(%rsp), %zmm0
vmovaps 0x7cdb7c(%rip), %zmm1 # 0x1e1d500
vcmpltps %zmm1, %zmm0, %k0
kmovw %k0, 0x13e(%rsp)
vmovaps 0x240(%rsp), %zmm1
vmovaps 0x1c0(%rsp), %zmm0
vmovaps %zmm1, 0x340(%rsp)
vmovaps %zmm0, 0x300(%rsp)
vmovaps 0x340(%rsp), %zmm0
vmovaps 0x300(%rsp), %zmm1
vsubps %zmm1, %zmm0, %zmm0
vmovaps %zmm0, 0xc0(%rsp)
vmovaps 0x140(%rsp), %zmm1
movw 0x13e(%rsp), %ax
vmovaps 0x1c0(%rsp), %zmm0
vmovaps %zmm1, 0xe00(%rsp)
movw %ax, 0xdfe(%rsp)
vmovaps %zmm1, 0xd80(%rsp)
vmovaps %zmm0, 0xd40(%rsp)
vmovaps 0xd80(%rsp), %zmm1
vmovaps 0xd40(%rsp), %zmm0
kmovw 0xdfe(%rsp), %k1
vmovaps %zmm1, 0xe80(%rsp)
vmovaps %zmm0, 0xe40(%rsp)
vmovaps 0xe80(%rsp), %zmm1
vmovaps 0xe40(%rsp), %zmm2
vmovaps 0xe00(%rsp), %zmm0
vsubps %zmm2, %zmm1, %zmm0 {%k1}
vmovaps %zmm0, 0x140(%rsp)
vmovaps 0xc0(%rsp), %zmm1
movw 0x13e(%rsp), %ax
vmovaps 0x240(%rsp), %zmm0
vmovaps %zmm1, 0x1400(%rsp)
movw %ax, 0x13fe(%rsp)
vmovaps %zmm1, 0x1380(%rsp)
vmovaps %zmm0, 0x1340(%rsp)
vmovaps 0x1380(%rsp), %zmm1
vmovaps 0x1340(%rsp), %zmm0
kmovw 0x13fe(%rsp), %k1
vmovaps %zmm1, 0x1480(%rsp)
vmovaps %zmm0, 0x1440(%rsp)
vmovaps 0x1480(%rsp), %zmm1
vmovaps 0x1440(%rsp), %zmm2
vmovaps 0x1400(%rsp), %zmm0
vaddps %zmm2, %zmm1, %zmm0 {%k1}
vmovaps %zmm0, 0x240(%rsp)
vmovaps 0x240(%rsp), %zmm0
vmovaps %zmm0, 0x500(%rsp)
vmovaps %zmm0, 0x4c0(%rsp)
vmovaps 0x500(%rsp), %zmm0
vmovaps 0x4c0(%rsp), %zmm1
vmulps %zmm1, %zmm0, %zmm0
vmovaps %zmm0, 0x80(%rsp)
vmovaps 0x7cda20(%rip), %zmm0 # 0x1e1d540
vmovaps %zmm0, 0x40(%rsp)
vmovaps 0x40(%rsp), %zmm2
vmovaps 0x240(%rsp), %zmm1
vmovaps 0x7cda3e(%rip), %zmm0 # 0x1e1d580
vmovaps %zmm2, 0xd00(%rsp)
vmovaps %zmm1, 0xcc0(%rsp)
vmovaps %zmm0, 0xc80(%rsp)
vmovaps 0xd00(%rsp), %zmm1
vmovaps 0xcc0(%rsp), %zmm0
vmovaps 0xc80(%rsp), %zmm2
vfmadd213ps %zmm2, %zmm1, %zmm0 # zmm0 = (zmm1 * zmm0) + zmm2
vmovaps %zmm0, 0x40(%rsp)
vmovaps 0x40(%rsp), %zmm2
vmovaps 0x240(%rsp), %zmm1
vmovaps 0x7cda26(%rip), %zmm0 # 0x1e1d5c0
vmovaps %zmm2, 0xc40(%rsp)
vmovaps %zmm1, 0xc00(%rsp)
vmovaps %zmm0, 0xbc0(%rsp)
vmovaps 0xc40(%rsp), %zmm1
vmovaps 0xc00(%rsp), %zmm0
vmovaps 0xbc0(%rsp), %zmm2
vfmadd213ps %zmm2, %zmm1, %zmm0 # zmm0 = (zmm1 * zmm0) + zmm2
vmovaps %zmm0, 0x40(%rsp)
vmovaps 0x40(%rsp), %zmm2
vmovaps 0x240(%rsp), %zmm1
vmovaps 0x7cda0e(%rip), %zmm0 # 0x1e1d600
vmovaps %zmm2, 0xb80(%rsp)
vmovaps %zmm1, 0xb40(%rsp)
vmovaps %zmm0, 0xb00(%rsp)
vmovaps 0xb80(%rsp), %zmm1
vmovaps 0xb40(%rsp), %zmm0
vmovaps 0xb00(%rsp), %zmm2
vfmadd213ps %zmm2, %zmm1, %zmm0 # zmm0 = (zmm1 * zmm0) + zmm2
vmovaps %zmm0, 0x40(%rsp)
vmovaps 0x40(%rsp), %zmm2
vmovaps 0x240(%rsp), %zmm1
vmovaps 0x7cd9f6(%rip), %zmm0 # 0x1e1d640
vmovaps %zmm2, 0xac0(%rsp)
vmovaps %zmm1, 0xa80(%rsp)
vmovaps %zmm0, 0xa40(%rsp)
vmovaps 0xac0(%rsp), %zmm1
vmovaps 0xa80(%rsp), %zmm0
vmovaps 0xa40(%rsp), %zmm2
vfmadd213ps %zmm2, %zmm1, %zmm0 # zmm0 = (zmm1 * zmm0) + zmm2
vmovaps %zmm0, 0x40(%rsp)
vmovaps 0x40(%rsp), %zmm2
vmovaps 0x240(%rsp), %zmm1
vmovaps 0x7cd9de(%rip), %zmm0 # 0x1e1d680
vmovaps %zmm2, 0xa00(%rsp)
vmovaps %zmm1, 0x9c0(%rsp)
vmovaps %zmm0, 0x980(%rsp)
vmovaps 0xa00(%rsp), %zmm1
vmovaps 0x9c0(%rsp), %zmm0
vmovaps 0x980(%rsp), %zmm2
vfmadd213ps %zmm2, %zmm1, %zmm0 # zmm0 = (zmm1 * zmm0) + zmm2
vmovaps %zmm0, 0x40(%rsp)
vmovaps 0x40(%rsp), %zmm2
vmovaps 0x240(%rsp), %zmm1
vmovaps 0x7cd9c6(%rip), %zmm0 # 0x1e1d6c0
vmovaps %zmm2, 0x940(%rsp)
vmovaps %zmm1, 0x900(%rsp)
vmovaps %zmm0, 0x8c0(%rsp)
vmovaps 0x940(%rsp), %zmm1
vmovaps 0x900(%rsp), %zmm0
vmovaps 0x8c0(%rsp), %zmm2
vfmadd213ps %zmm2, %zmm1, %zmm0 # zmm0 = (zmm1 * zmm0) + zmm2
vmovaps %zmm0, 0x40(%rsp)
vmovaps 0x40(%rsp), %zmm2
vmovaps 0x240(%rsp), %zmm1
vmovaps 0x7cd9ae(%rip), %zmm0 # 0x1e1d700
vmovaps %zmm2, 0x880(%rsp)
vmovaps %zmm1, 0x840(%rsp)
vmovaps %zmm0, 0x800(%rsp)
vmovaps 0x880(%rsp), %zmm1
vmovaps 0x840(%rsp), %zmm0
vmovaps 0x800(%rsp), %zmm2
vfmadd213ps %zmm2, %zmm1, %zmm0 # zmm0 = (zmm1 * zmm0) + zmm2
vmovaps %zmm0, 0x40(%rsp)
vmovaps 0x40(%rsp), %zmm2
vmovaps 0x240(%rsp), %zmm1
vmovaps 0x7cd996(%rip), %zmm0 # 0x1e1d740
vmovaps %zmm2, 0x7c0(%rsp)
vmovaps %zmm1, 0x780(%rsp)
vmovaps %zmm0, 0x740(%rsp)
vmovaps 0x7c0(%rsp), %zmm1
vmovaps 0x780(%rsp), %zmm0
vmovaps 0x740(%rsp), %zmm2
vfmadd213ps %zmm2, %zmm1, %zmm0 # zmm0 = (zmm1 * zmm0) + zmm2
vmovaps %zmm0, 0x40(%rsp)
vmovaps 0x40(%rsp), %zmm1
vmovaps 0x240(%rsp), %zmm0
vmovaps %zmm1, 0x480(%rsp)
vmovaps %zmm0, 0x440(%rsp)
vmovaps 0x480(%rsp), %zmm0
vmovaps 0x440(%rsp), %zmm1
vmulps %zmm1, %zmm0, %zmm0
vmovaps %zmm0, 0x40(%rsp)
vmovaps 0x40(%rsp), %zmm1
vmovaps 0x80(%rsp), %zmm0
vmovaps %zmm1, 0x400(%rsp)
vmovaps %zmm0, 0x3c0(%rsp)
vmovaps 0x400(%rsp), %zmm0
vmovaps 0x3c0(%rsp), %zmm1
vmulps %zmm1, %zmm0, %zmm0
vmovaps %zmm0, 0x40(%rsp)
vmovaps 0x140(%rsp), %zmm2
vmovaps 0x7cd90a(%rip), %zmm1 # 0x1e1d780
vmovaps 0x40(%rsp), %zmm0
vmovaps %zmm2, 0x700(%rsp)
vmovaps %zmm1, 0x6c0(%rsp)
vmovaps %zmm0, 0x680(%rsp)
vmovaps 0x700(%rsp), %zmm1
vmovaps 0x6c0(%rsp), %zmm0
vmovaps 0x680(%rsp), %zmm2
vfmadd213ps %zmm2, %zmm1, %zmm0 # zmm0 = (zmm1 * zmm0) + zmm2
vmovaps %zmm0, 0x40(%rsp)
vmovaps 0x80(%rsp), %zmm2
vmovaps 0x7cd072(%rip), %zmm1 # 0x1e1cf40
vmovaps 0x40(%rsp), %zmm0
vmovaps %zmm2, 0xf40(%rsp)
vmovaps %zmm1, 0xf00(%rsp)
vmovaps %zmm0, 0xec0(%rsp)
vmovaps 0xf40(%rsp), %zmm1
vmovaps 0xf00(%rsp), %zmm0
vmovaps 0xec0(%rsp), %zmm2
vfnmadd213ps %zmm2, %zmm1, %zmm0 # zmm0 = -(zmm1 * zmm0) + zmm2
vmovaps %zmm0, 0x40(%rsp)
vmovaps 0x240(%rsp), %zmm1
vmovaps 0x40(%rsp), %zmm0
vmovaps %zmm1, 0xfc0(%rsp)
vmovaps %zmm0, 0xf80(%rsp)
vmovaps 0xfc0(%rsp), %zmm0
vmovaps 0xf80(%rsp), %zmm1
vaddps %zmm1, %zmm0, %zmm0
vmovaps %zmm0, 0x240(%rsp)
vmovaps 0x140(%rsp), %zmm2
vmovaps 0x7cd85c(%rip), %zmm1 # 0x1e1d7c0
vmovaps 0x240(%rsp), %zmm0
vmovaps %zmm2, 0x640(%rsp)
vmovaps %zmm1, 0x600(%rsp)
vmovaps %zmm0, 0x5c0(%rsp)
vmovaps 0x640(%rsp), %zmm1
vmovaps 0x600(%rsp), %zmm0
vmovaps 0x5c0(%rsp), %zmm2
vfmadd213ps %zmm2, %zmm1, %zmm0 # zmm0 = (zmm1 * zmm0) + zmm2
vmovaps %zmm0, 0x240(%rsp)
vmovaps 0x240(%rsp), %zmm1
movw 0x1be(%rsp), %ax
movw %ax, 0x14ee(%rsp)
movw 0x14ee(%rsp), %cx
movw %cx, %ax
kmovd %eax, %k0
vpmovm2d %k0, %zmm0
vmovdqa64 %zmm0, 0x280(%rsp)
vmovdqa64 0x280(%rsp), %zmm0
vmovaps %zmm1, 0x11c0(%rsp)
vmovaps %zmm0, 0x1180(%rsp)
vmovaps 0x11c0(%rsp), %zmm0
vmovaps 0x1180(%rsp), %zmm1
vpord %zmm1, %zmm0, %zmm0
vmovaps %zmm0, 0x40(%rsp)
vmovaps 0x40(%rsp), %zmm0
movq %rbp, %rsp
popq %rbp
retq
nopw %cs:(%rax,%rax)
nopl (%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,286 |
ncnn::UnaryOp_x86_avx512_functor::unary_op_log::func_pack8(float vector[8] const&) const
|
__m256 func_pack8(const __m256& x) const
{
return log256_ps(x);
}
|
pushq %rbp
movq %rsp, %rbp
andq $-0x20, %rsp
subq $0xe80, %rsp # imm = 0xE80
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
movq 0x10(%rsp), %rax
vmovaps (%rax), %ymm0
vmovaps %ymm0, 0x120(%rsp)
vbroadcastss 0x7b8fa0(%rip), %ymm0 # 0x1e09004
vmovaps %ymm0, 0xe0(%rsp)
vmovaps 0x120(%rsp), %ymm0
vpxor %xmm1, %xmm1, %xmm1
vmovaps %ymm1, 0x200(%rsp)
vmovaps 0x200(%rsp), %ymm1
vcmpleps %ymm1, %ymm0, %ymm0
vmovdqa %ymm0, 0xc0(%rsp)
vmovaps 0x120(%rsp), %ymm0
vmovaps %ymm0, 0x160(%rsp)
vbroadcastss 0x7bbd33(%rip), %ymm0 # 0x1e0bde8
vmovaps %ymm0, 0x140(%rsp)
vmovaps 0x160(%rsp), %ymm0
vmovaps 0x140(%rsp), %ymm1
vmaxps %ymm1, %ymm0, %ymm0
vmovaps %ymm0, 0x120(%rsp)
vmovaps 0x120(%rsp), %ymm0
vmovaps %ymm0, 0xce0(%rsp)
vmovaps 0xce0(%rsp), %ymm0
vmovaps %ymm0, 0xcc0(%rsp)
movl $0x17, 0xcbc(%rsp)
vmovdqa 0xcc0(%rsp), %ymm0
movl 0xcbc(%rsp), %eax
vmovdqa %ymm0, 0xe00(%rsp)
movl %eax, 0xdfc(%rsp)
vmovdqa 0xe00(%rsp), %ymm0
vmovd 0xdfc(%rsp), %xmm1
vpsrld %xmm1, %ymm0, %ymm0
vmovdqa %ymm0, 0x100(%rsp)
vmovaps 0x120(%rsp), %ymm0
vmovaps %ymm0, 0x740(%rsp)
vbroadcastss 0x7bbc86(%rip), %ymm0 # 0x1e0bdec
vmovaps %ymm0, 0x720(%rsp)
vmovdqa 0x740(%rsp), %ymm0
vmovdqa 0x720(%rsp), %ymm1
vpand %ymm1, %ymm0, %ymm0
vmovdqa %ymm0, 0x120(%rsp)
vmovaps 0x120(%rsp), %ymm0
vmovaps %ymm0, 0xd60(%rsp)
vbroadcastss 0x7b8e73(%rip), %ymm1 # 0x1e0901c
vmovaps %ymm1, 0xd40(%rsp)
vmovdqa 0xd60(%rsp), %ymm0
vmovdqa 0xd40(%rsp), %ymm2
vpor %ymm2, %ymm0, %ymm0
vmovdqa %ymm0, 0x120(%rsp)
vmovdqa 0x100(%rsp), %ymm0
vmovdqa %ymm0, 0xda0(%rsp)
vpbroadcastq 0x7bbc2c(%rip), %ymm0 # 0x1e0be18
vmovdqa %ymm0, 0xd80(%rsp)
vmovdqa 0xda0(%rsp), %ymm2
vmovdqa 0xd80(%rsp), %ymm0
vmovdqa %ymm2, 0xe40(%rsp)
vmovdqa %ymm0, 0xe20(%rsp)
vmovdqa 0xe40(%rsp), %ymm0
vmovdqa 0xe20(%rsp), %ymm2
vpsubd %ymm2, %ymm0, %ymm0
vmovdqa %ymm0, 0x100(%rsp)
vmovdqa 0x100(%rsp), %ymm0
vmovdqa %ymm0, 0xdc0(%rsp)
vcvtdq2ps 0xdc0(%rsp), %ymm0
vmovaps %ymm0, 0xa0(%rsp)
vmovaps 0xa0(%rsp), %ymm2
vmovaps 0xe0(%rsp), %ymm0
vmovaps %ymm2, 0x860(%rsp)
vmovaps %ymm0, 0x840(%rsp)
vmovaps 0x860(%rsp), %ymm0
vmovaps 0x840(%rsp), %ymm2
vaddps %ymm2, %ymm0, %ymm0
vmovaps %ymm0, 0xa0(%rsp)
vmovaps 0x120(%rsp), %ymm0
vbroadcastss 0x7bbb3f(%rip), %ymm2 # 0x1e0bdf0
vcmpltps %ymm2, %ymm0, %ymm0
vmovdqa %ymm0, 0x80(%rsp)
vmovaps 0x120(%rsp), %ymm2
vmovaps 0x80(%rsp), %ymm0
vmovaps %ymm2, 0x700(%rsp)
vmovaps %ymm0, 0x6e0(%rsp)
vmovdqa 0x700(%rsp), %ymm0
vmovdqa 0x6e0(%rsp), %ymm2
vpand %ymm2, %ymm0, %ymm0
vmovdqa %ymm0, 0x60(%rsp)
vmovaps 0x120(%rsp), %ymm2
vmovaps 0xe0(%rsp), %ymm0
vmovaps %ymm2, 0x1e0(%rsp)
vmovaps %ymm0, 0x1c0(%rsp)
vmovaps 0x1e0(%rsp), %ymm0
vmovaps 0x1c0(%rsp), %ymm2
vsubps %ymm2, %ymm0, %ymm0
vmovaps %ymm0, 0x120(%rsp)
vmovaps 0xa0(%rsp), %ymm2
vmovaps 0xe0(%rsp), %ymm3
vmovaps 0x80(%rsp), %ymm0
vmovaps %ymm3, 0x6c0(%rsp)
vmovaps %ymm0, 0x6a0(%rsp)
vmovdqa 0x6c0(%rsp), %ymm0
vmovdqa 0x6a0(%rsp), %ymm3
vpand %ymm3, %ymm0, %ymm0
vmovaps %ymm2, 0x1a0(%rsp)
vmovdqa %ymm0, 0x180(%rsp)
vmovaps 0x1a0(%rsp), %ymm0
vmovaps 0x180(%rsp), %ymm2
vsubps %ymm2, %ymm0, %ymm0
vmovaps %ymm0, 0xa0(%rsp)
vmovaps 0x120(%rsp), %ymm2
vmovaps 0x60(%rsp), %ymm0
vmovaps %ymm2, 0x820(%rsp)
vmovaps %ymm0, 0x800(%rsp)
vmovaps 0x820(%rsp), %ymm0
vmovaps 0x800(%rsp), %ymm2
vaddps %ymm2, %ymm0, %ymm0
vmovaps %ymm0, 0x120(%rsp)
vmovaps 0x120(%rsp), %ymm0
vmovaps %ymm0, 0x2c0(%rsp)
vmovaps %ymm0, 0x2a0(%rsp)
vmovaps 0x2c0(%rsp), %ymm0
vmovaps 0x2a0(%rsp), %ymm2
vmulps %ymm2, %ymm0, %ymm0
vmovaps %ymm0, 0x40(%rsp)
vbroadcastss 0x7bb9be(%rip), %ymm0 # 0x1e0bdf4
vmovaps %ymm0, 0x20(%rsp)
vmovaps 0x20(%rsp), %ymm2
vmovaps 0x120(%rsp), %ymm0
vmovaps %ymm2, 0x680(%rsp)
vmovaps %ymm0, 0x660(%rsp)
vbroadcastss 0x7bb992(%rip), %ymm0 # 0x1e0bdf8
vmovaps %ymm0, 0x640(%rsp)
vmovaps 0x680(%rsp), %ymm3
vmovaps 0x660(%rsp), %ymm2
vmovaps 0x640(%rsp), %ymm0
vmovaps %ymm3, 0x8c0(%rsp)
vmovaps %ymm2, 0x8a0(%rsp)
vmovaps %ymm0, 0x880(%rsp)
vmovaps 0x8c0(%rsp), %ymm2
vmovaps 0x8a0(%rsp), %ymm0
vmovaps 0x880(%rsp), %ymm3
vfmadd213ps %ymm3, %ymm2, %ymm0 # ymm0 = (ymm2 * ymm0) + ymm3
vmovaps %ymm0, 0x20(%rsp)
vmovaps 0x20(%rsp), %ymm2
vmovaps 0x120(%rsp), %ymm0
vmovaps %ymm2, 0x620(%rsp)
vmovaps %ymm0, 0x600(%rsp)
vbroadcastss 0x7bb907(%rip), %ymm0 # 0x1e0bdfc
vmovaps %ymm0, 0x5e0(%rsp)
vmovaps 0x620(%rsp), %ymm3
vmovaps 0x600(%rsp), %ymm2
vmovaps 0x5e0(%rsp), %ymm0
vmovaps %ymm3, 0x920(%rsp)
vmovaps %ymm2, 0x900(%rsp)
vmovaps %ymm0, 0x8e0(%rsp)
vmovaps 0x920(%rsp), %ymm2
vmovaps 0x900(%rsp), %ymm0
vmovaps 0x8e0(%rsp), %ymm3
vfmadd213ps %ymm3, %ymm2, %ymm0 # ymm0 = (ymm2 * ymm0) + ymm3
vmovaps %ymm0, 0x20(%rsp)
vmovaps 0x20(%rsp), %ymm2
vmovaps 0x120(%rsp), %ymm0
vmovaps %ymm2, 0x5c0(%rsp)
vmovaps %ymm0, 0x5a0(%rsp)
vbroadcastss 0x7bb87c(%rip), %ymm0 # 0x1e0be00
vmovaps %ymm0, 0x580(%rsp)
vmovaps 0x5c0(%rsp), %ymm3
vmovaps 0x5a0(%rsp), %ymm2
vmovaps 0x580(%rsp), %ymm0
vmovaps %ymm3, 0x980(%rsp)
vmovaps %ymm2, 0x960(%rsp)
vmovaps %ymm0, 0x940(%rsp)
vmovaps 0x980(%rsp), %ymm2
vmovaps 0x960(%rsp), %ymm0
vmovaps 0x940(%rsp), %ymm3
vfmadd213ps %ymm3, %ymm2, %ymm0 # ymm0 = (ymm2 * ymm0) + ymm3
vmovaps %ymm0, 0x20(%rsp)
vmovaps 0x20(%rsp), %ymm2
vmovaps 0x120(%rsp), %ymm0
vmovaps %ymm2, 0x560(%rsp)
vmovaps %ymm0, 0x540(%rsp)
vbroadcastss 0x7bb7f1(%rip), %ymm0 # 0x1e0be04
vmovaps %ymm0, 0x520(%rsp)
vmovaps 0x560(%rsp), %ymm3
vmovaps 0x540(%rsp), %ymm2
vmovaps 0x520(%rsp), %ymm0
vmovaps %ymm3, 0x9e0(%rsp)
vmovaps %ymm2, 0x9c0(%rsp)
vmovaps %ymm0, 0x9a0(%rsp)
vmovaps 0x9e0(%rsp), %ymm2
vmovaps 0x9c0(%rsp), %ymm0
vmovaps 0x9a0(%rsp), %ymm3
vfmadd213ps %ymm3, %ymm2, %ymm0 # ymm0 = (ymm2 * ymm0) + ymm3
vmovaps %ymm0, 0x20(%rsp)
vmovaps 0x20(%rsp), %ymm2
vmovaps 0x120(%rsp), %ymm0
vmovaps %ymm2, 0x500(%rsp)
vmovaps %ymm0, 0x4e0(%rsp)
vbroadcastss 0x7bb766(%rip), %ymm0 # 0x1e0be08
vmovaps %ymm0, 0x4c0(%rsp)
vmovaps 0x500(%rsp), %ymm3
vmovaps 0x4e0(%rsp), %ymm2
vmovaps 0x4c0(%rsp), %ymm0
vmovaps %ymm3, 0xa40(%rsp)
vmovaps %ymm2, 0xa20(%rsp)
vmovaps %ymm0, 0xa00(%rsp)
vmovaps 0xa40(%rsp), %ymm2
vmovaps 0xa20(%rsp), %ymm0
vmovaps 0xa00(%rsp), %ymm3
vfmadd213ps %ymm3, %ymm2, %ymm0 # ymm0 = (ymm2 * ymm0) + ymm3
vmovaps %ymm0, 0x20(%rsp)
vmovaps 0x20(%rsp), %ymm2
vmovaps 0x120(%rsp), %ymm0
vmovaps %ymm2, 0x4a0(%rsp)
vmovaps %ymm0, 0x480(%rsp)
vbroadcastss 0x7bb6db(%rip), %ymm0 # 0x1e0be0c
vmovaps %ymm0, 0x460(%rsp)
vmovaps 0x4a0(%rsp), %ymm3
vmovaps 0x480(%rsp), %ymm2
vmovaps 0x460(%rsp), %ymm0
vmovaps %ymm3, 0xaa0(%rsp)
vmovaps %ymm2, 0xa80(%rsp)
vmovaps %ymm0, 0xa60(%rsp)
vmovaps 0xaa0(%rsp), %ymm2
vmovaps 0xa80(%rsp), %ymm0
vmovaps 0xa60(%rsp), %ymm3
vfmadd213ps %ymm3, %ymm2, %ymm0 # ymm0 = (ymm2 * ymm0) + ymm3
vmovaps %ymm0, 0x20(%rsp)
vmovaps 0x20(%rsp), %ymm2
vmovaps 0x120(%rsp), %ymm0
vmovaps %ymm2, 0x440(%rsp)
vmovaps %ymm0, 0x420(%rsp)
vbroadcastss 0x7bb650(%rip), %ymm0 # 0x1e0be10
vmovaps %ymm0, 0x400(%rsp)
vmovaps 0x440(%rsp), %ymm3
vmovaps 0x420(%rsp), %ymm2
vmovaps 0x400(%rsp), %ymm0
vmovaps %ymm3, 0xb00(%rsp)
vmovaps %ymm2, 0xae0(%rsp)
vmovaps %ymm0, 0xac0(%rsp)
vmovaps 0xb00(%rsp), %ymm2
vmovaps 0xae0(%rsp), %ymm0
vmovaps 0xac0(%rsp), %ymm3
vfmadd213ps %ymm3, %ymm2, %ymm0 # ymm0 = (ymm2 * ymm0) + ymm3
vmovaps %ymm0, 0x20(%rsp)
vmovaps 0x20(%rsp), %ymm2
vmovaps 0x120(%rsp), %ymm0
vmovaps %ymm2, 0x3e0(%rsp)
vmovaps %ymm0, 0x3c0(%rsp)
vbroadcastss 0x7bb5c5(%rip), %ymm0 # 0x1e0be14
vmovaps %ymm0, 0x3a0(%rsp)
vmovaps 0x3e0(%rsp), %ymm3
vmovaps 0x3c0(%rsp), %ymm2
vmovaps 0x3a0(%rsp), %ymm0
vmovaps %ymm3, 0xb60(%rsp)
vmovaps %ymm2, 0xb40(%rsp)
vmovaps %ymm0, 0xb20(%rsp)
vmovaps 0xb60(%rsp), %ymm2
vmovaps 0xb40(%rsp), %ymm0
vmovaps 0xb20(%rsp), %ymm3
vfmadd213ps %ymm3, %ymm2, %ymm0 # ymm0 = (ymm2 * ymm0) + ymm3
vmovaps %ymm0, 0x20(%rsp)
vmovaps 0x20(%rsp), %ymm2
vmovaps 0x120(%rsp), %ymm0
vmovaps %ymm2, 0x280(%rsp)
vmovaps %ymm0, 0x260(%rsp)
vmovaps 0x280(%rsp), %ymm0
vmovaps 0x260(%rsp), %ymm2
vmulps %ymm2, %ymm0, %ymm0
vmovaps %ymm0, 0x20(%rsp)
vmovaps 0x20(%rsp), %ymm2
vmovaps 0x40(%rsp), %ymm0
vmovaps %ymm2, 0x240(%rsp)
vmovaps %ymm0, 0x220(%rsp)
vmovaps 0x240(%rsp), %ymm0
vmovaps 0x220(%rsp), %ymm2
vmulps %ymm2, %ymm0, %ymm0
vmovaps %ymm0, 0x20(%rsp)
vmovaps 0xa0(%rsp), %ymm2
vmovaps 0x20(%rsp), %ymm0
vmovaps %ymm2, 0x380(%rsp)
vbroadcastss 0x7bb484(%rip), %ymm2 # 0x1e0bdd0
vmovaps %ymm2, 0x360(%rsp)
vmovaps %ymm0, 0x340(%rsp)
vmovaps 0x380(%rsp), %ymm3
vmovaps 0x360(%rsp), %ymm2
vmovaps 0x340(%rsp), %ymm0
vmovaps %ymm3, 0xbc0(%rsp)
vmovaps %ymm2, 0xba0(%rsp)
vmovaps %ymm0, 0xb80(%rsp)
vmovaps 0xbc0(%rsp), %ymm2
vmovaps 0xba0(%rsp), %ymm0
vmovaps 0xb80(%rsp), %ymm3
vfmadd213ps %ymm3, %ymm2, %ymm0 # ymm0 = (ymm2 * ymm0) + ymm3
vmovaps %ymm0, 0x20(%rsp)
vmovaps 0x40(%rsp), %ymm2
vmovaps 0x20(%rsp), %ymm0
vmovaps %ymm2, 0x7a0(%rsp)
vmovaps %ymm1, 0x780(%rsp)
vmovaps %ymm0, 0x760(%rsp)
vmovaps 0x7a0(%rsp), %ymm2
vmovaps 0x780(%rsp), %ymm1
vmovaps 0x760(%rsp), %ymm0
vmovaps %ymm2, 0xc80(%rsp)
vmovaps %ymm1, 0xc60(%rsp)
vmovaps %ymm0, 0xc40(%rsp)
vmovaps 0xc80(%rsp), %ymm1
vmovaps 0xc60(%rsp), %ymm0
vmovaps 0xc40(%rsp), %ymm2
vfnmadd213ps %ymm2, %ymm1, %ymm0 # ymm0 = -(ymm1 * ymm0) + ymm2
vmovaps %ymm0, 0x20(%rsp)
vmovaps 0x120(%rsp), %ymm1
vmovaps 0x20(%rsp), %ymm0
vmovaps %ymm1, 0x7e0(%rsp)
vmovaps %ymm0, 0x7c0(%rsp)
vmovaps 0x7e0(%rsp), %ymm0
vmovaps 0x7c0(%rsp), %ymm1
vaddps %ymm1, %ymm0, %ymm0
vmovaps %ymm0, 0x120(%rsp)
vmovaps 0xa0(%rsp), %ymm1
vmovaps 0x120(%rsp), %ymm0
vmovaps %ymm1, 0x320(%rsp)
vbroadcastss 0x7bb32b(%rip), %ymm1 # 0x1e0bdcc
vmovaps %ymm1, 0x300(%rsp)
vmovaps %ymm0, 0x2e0(%rsp)
vmovaps 0x320(%rsp), %ymm2
vmovaps 0x300(%rsp), %ymm1
vmovaps 0x2e0(%rsp), %ymm0
vmovaps %ymm2, 0xc20(%rsp)
vmovaps %ymm1, 0xc00(%rsp)
vmovaps %ymm0, 0xbe0(%rsp)
vmovaps 0xc20(%rsp), %ymm1
vmovaps 0xc00(%rsp), %ymm0
vmovaps 0xbe0(%rsp), %ymm2
vfmadd213ps %ymm2, %ymm1, %ymm0 # ymm0 = (ymm1 * ymm0) + ymm2
vmovaps %ymm0, 0x120(%rsp)
vmovaps 0x120(%rsp), %ymm1
vmovaps 0xc0(%rsp), %ymm0
vmovaps %ymm1, 0xd20(%rsp)
vmovaps %ymm0, 0xd00(%rsp)
vmovaps 0xd20(%rsp), %ymm0
vmovaps 0xd00(%rsp), %ymm1
vpor %ymm1, %ymm0, %ymm0
vmovaps %ymm0, 0x20(%rsp)
vmovaps 0x20(%rsp), %ymm0
movq %rbp, %rsp
popq %rbp
retq
nopl (%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,287 |
ncnn::UnaryOp_x86_avx512_functor::unary_op_log::func_pack4(float vector[4] const&) const
|
__m128 func_pack4(const __m128& x) const
{
return log_ps(x);
}
|
subq $0x6b8, %rsp # imm = 0x6B8
movq %rdi, -0x78(%rsp)
movq %rsi, -0x80(%rsp)
movq -0x80(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm0, 0x10(%rsp)
vbroadcastss 0x7b847b(%rip), %xmm0 # 0x1e09004
vmovaps %xmm0, -0x10(%rsp)
vmovaps 0x10(%rsp), %xmm1
vpxor %xmm0, %xmm0, %xmm0
vmovaps %xmm0, 0x60(%rsp)
vmovaps 0x60(%rsp), %xmm0
vmovaps %xmm1, 0x5f0(%rsp)
vmovaps %xmm0, 0x5e0(%rsp)
vmovaps 0x5f0(%rsp), %xmm0
vmovaps 0x5e0(%rsp), %xmm1
vcmpleps %xmm1, %xmm0, %xmm0
vmovaps %xmm0, -0x20(%rsp)
vmovaps 0x10(%rsp), %xmm0
vmovaps %xmm0, 0xe0(%rsp)
vbroadcastss 0x7bb1fc(%rip), %xmm0 # 0x1e0bde8
vmovaps %xmm0, 0xd0(%rsp)
vmovaps 0xe0(%rsp), %xmm0
vmovaps 0xd0(%rsp), %xmm1
vmaxps %xmm1, %xmm0, %xmm0
vmovaps %xmm0, 0x10(%rsp)
vmovaps 0x10(%rsp), %xmm0
vmovaps %xmm0, 0x620(%rsp)
vmovaps 0x620(%rsp), %xmm0
vmovaps %xmm0, 0x610(%rsp)
movl $0x17, 0x60c(%rsp)
vmovdqa 0x610(%rsp), %xmm0
vmovd 0x60c(%rsp), %xmm1
vpsrld %xmm1, %xmm0, %xmm0
vmovdqa %xmm0, (%rsp)
vmovaps 0x10(%rsp), %xmm0
vmovaps %xmm0, 0x1b0(%rsp)
vbroadcastss 0x7bb17c(%rip), %xmm0 # 0x1e0bdec
vmovaps %xmm0, 0x1a0(%rsp)
vmovdqa 0x1b0(%rsp), %xmm0
vmovdqa 0x1a0(%rsp), %xmm1
vpand %xmm1, %xmm0, %xmm0
vmovdqa %xmm0, 0x10(%rsp)
vmovaps 0x10(%rsp), %xmm0
vmovaps %xmm0, 0x660(%rsp)
vbroadcastss 0x7b836f(%rip), %xmm1 # 0x1e0901c
vmovaps %xmm1, 0x650(%rsp)
vmovdqa 0x660(%rsp), %xmm0
vmovdqa 0x650(%rsp), %xmm2
vpor %xmm2, %xmm0, %xmm0
vmovdqa %xmm0, 0x10(%rsp)
vmovdqa (%rsp), %xmm0
vmovdqa %xmm0, 0x680(%rsp)
vpbroadcastq 0x7bb12f(%rip), %xmm0 # 0x1e0be18
vmovdqa %xmm0, 0x670(%rsp)
vmovdqa 0x680(%rsp), %xmm0
vmovdqa 0x670(%rsp), %xmm2
vpsubd %xmm2, %xmm0, %xmm0
vmovdqa %xmm0, (%rsp)
vmovdqa (%rsp), %xmm0
vmovdqa %xmm0, 0x150(%rsp)
vcvtdq2ps 0x150(%rsp), %xmm0
vmovaps %xmm0, -0x30(%rsp)
vmovaps -0x30(%rsp), %xmm2
vmovaps -0x10(%rsp), %xmm0
vmovaps %xmm2, 0x140(%rsp)
vmovaps %xmm0, 0x130(%rsp)
vmovaps 0x140(%rsp), %xmm0
vmovaps 0x130(%rsp), %xmm2
vaddps %xmm2, %xmm0, %xmm0
vmovaps %xmm0, -0x30(%rsp)
vmovaps 0x10(%rsp), %xmm0
vmovaps %xmm0, 0x6a0(%rsp)
vbroadcastss 0x7bb074(%rip), %xmm0 # 0x1e0bdf0
vmovaps %xmm0, 0x690(%rsp)
vmovaps 0x6a0(%rsp), %xmm0
vmovaps 0x690(%rsp), %xmm2
vcmpltps %xmm2, %xmm0, %xmm0
vmovaps %xmm0, -0x40(%rsp)
vmovaps 0x10(%rsp), %xmm2
vmovaps -0x40(%rsp), %xmm0
vmovaps %xmm2, 0x190(%rsp)
vmovaps %xmm0, 0x180(%rsp)
vmovdqa 0x190(%rsp), %xmm0
vmovdqa 0x180(%rsp), %xmm2
vpand %xmm2, %xmm0, %xmm0
vmovdqa %xmm0, -0x50(%rsp)
vmovaps 0x10(%rsp), %xmm2
vmovaps -0x10(%rsp), %xmm0
vmovaps %xmm2, 0x50(%rsp)
vmovaps %xmm0, 0x40(%rsp)
vmovaps 0x50(%rsp), %xmm0
vmovaps 0x40(%rsp), %xmm2
vsubps %xmm2, %xmm0, %xmm0
vmovaps %xmm0, 0x10(%rsp)
vmovaps -0x30(%rsp), %xmm2
vmovaps -0x10(%rsp), %xmm3
vmovaps -0x40(%rsp), %xmm0
vmovaps %xmm3, 0x170(%rsp)
vmovaps %xmm0, 0x160(%rsp)
vmovdqa 0x170(%rsp), %xmm0
vmovdqa 0x160(%rsp), %xmm3
vpand %xmm3, %xmm0, %xmm0
vmovaps %xmm2, 0x30(%rsp)
vmovdqa %xmm0, 0x20(%rsp)
vmovaps 0x30(%rsp), %xmm0
vmovaps 0x20(%rsp), %xmm2
vsubps %xmm2, %xmm0, %xmm0
vmovaps %xmm0, -0x30(%rsp)
vmovaps 0x10(%rsp), %xmm2
vmovaps -0x50(%rsp), %xmm0
vmovaps %xmm2, 0x120(%rsp)
vmovaps %xmm0, 0x110(%rsp)
vmovaps 0x120(%rsp), %xmm0
vmovaps 0x110(%rsp), %xmm2
vaddps %xmm2, %xmm0, %xmm0
vmovaps %xmm0, 0x10(%rsp)
vmovaps 0x10(%rsp), %xmm0
vmovaps %xmm0, 0xc0(%rsp)
vmovaps %xmm0, 0xb0(%rsp)
vmovaps 0xc0(%rsp), %xmm0
vmovaps 0xb0(%rsp), %xmm2
vmulps %xmm2, %xmm0, %xmm0
vmovaps %xmm0, -0x60(%rsp)
vbroadcastss 0x7baf17(%rip), %xmm0 # 0x1e0bdf4
vmovaps %xmm0, -0x70(%rsp)
vmovaps -0x70(%rsp), %xmm2
vmovaps 0x10(%rsp), %xmm0
vmovaps %xmm2, 0x3c0(%rsp)
vmovaps %xmm0, 0x3b0(%rsp)
vbroadcastss 0x7baeee(%rip), %xmm0 # 0x1e0bdf8
vmovaps %xmm0, 0x3a0(%rsp)
vmovaps 0x3c0(%rsp), %xmm3
vmovaps 0x3b0(%rsp), %xmm2
vmovaps 0x3a0(%rsp), %xmm0
vmovaps %xmm3, 0x420(%rsp)
vmovaps %xmm2, 0x410(%rsp)
vmovaps %xmm0, 0x400(%rsp)
vmovaps 0x420(%rsp), %xmm2
vmovaps 0x410(%rsp), %xmm0
vmovaps 0x400(%rsp), %xmm3
vfmadd213ps %xmm3, %xmm2, %xmm0 # xmm0 = (xmm2 * xmm0) + xmm3
vmovaps %xmm0, -0x70(%rsp)
vmovaps -0x70(%rsp), %xmm2
vmovaps 0x10(%rsp), %xmm0
vmovaps %xmm2, 0x390(%rsp)
vmovaps %xmm0, 0x380(%rsp)
vbroadcastss 0x7bae66(%rip), %xmm0 # 0x1e0bdfc
vmovaps %xmm0, 0x370(%rsp)
vmovaps 0x390(%rsp), %xmm3
vmovaps 0x380(%rsp), %xmm2
vmovaps 0x370(%rsp), %xmm0
vmovaps %xmm3, 0x450(%rsp)
vmovaps %xmm2, 0x440(%rsp)
vmovaps %xmm0, 0x430(%rsp)
vmovaps 0x450(%rsp), %xmm2
vmovaps 0x440(%rsp), %xmm0
vmovaps 0x430(%rsp), %xmm3
vfmadd213ps %xmm3, %xmm2, %xmm0 # xmm0 = (xmm2 * xmm0) + xmm3
vmovaps %xmm0, -0x70(%rsp)
vmovaps -0x70(%rsp), %xmm2
vmovaps 0x10(%rsp), %xmm0
vmovaps %xmm2, 0x360(%rsp)
vmovaps %xmm0, 0x350(%rsp)
vbroadcastss 0x7badde(%rip), %xmm0 # 0x1e0be00
vmovaps %xmm0, 0x340(%rsp)
vmovaps 0x360(%rsp), %xmm3
vmovaps 0x350(%rsp), %xmm2
vmovaps 0x340(%rsp), %xmm0
vmovaps %xmm3, 0x480(%rsp)
vmovaps %xmm2, 0x470(%rsp)
vmovaps %xmm0, 0x460(%rsp)
vmovaps 0x480(%rsp), %xmm2
vmovaps 0x470(%rsp), %xmm0
vmovaps 0x460(%rsp), %xmm3
vfmadd213ps %xmm3, %xmm2, %xmm0 # xmm0 = (xmm2 * xmm0) + xmm3
vmovaps %xmm0, -0x70(%rsp)
vmovaps -0x70(%rsp), %xmm2
vmovaps 0x10(%rsp), %xmm0
vmovaps %xmm2, 0x330(%rsp)
vmovaps %xmm0, 0x320(%rsp)
vbroadcastss 0x7bad56(%rip), %xmm0 # 0x1e0be04
vmovaps %xmm0, 0x310(%rsp)
vmovaps 0x330(%rsp), %xmm3
vmovaps 0x320(%rsp), %xmm2
vmovaps 0x310(%rsp), %xmm0
vmovaps %xmm3, 0x4b0(%rsp)
vmovaps %xmm2, 0x4a0(%rsp)
vmovaps %xmm0, 0x490(%rsp)
vmovaps 0x4b0(%rsp), %xmm2
vmovaps 0x4a0(%rsp), %xmm0
vmovaps 0x490(%rsp), %xmm3
vfmadd213ps %xmm3, %xmm2, %xmm0 # xmm0 = (xmm2 * xmm0) + xmm3
vmovaps %xmm0, -0x70(%rsp)
vmovaps -0x70(%rsp), %xmm2
vmovaps 0x10(%rsp), %xmm0
vmovaps %xmm2, 0x300(%rsp)
vmovaps %xmm0, 0x2f0(%rsp)
vbroadcastss 0x7bacce(%rip), %xmm0 # 0x1e0be08
vmovaps %xmm0, 0x2e0(%rsp)
vmovaps 0x300(%rsp), %xmm3
vmovaps 0x2f0(%rsp), %xmm2
vmovaps 0x2e0(%rsp), %xmm0
vmovaps %xmm3, 0x4e0(%rsp)
vmovaps %xmm2, 0x4d0(%rsp)
vmovaps %xmm0, 0x4c0(%rsp)
vmovaps 0x4e0(%rsp), %xmm2
vmovaps 0x4d0(%rsp), %xmm0
vmovaps 0x4c0(%rsp), %xmm3
vfmadd213ps %xmm3, %xmm2, %xmm0 # xmm0 = (xmm2 * xmm0) + xmm3
vmovaps %xmm0, -0x70(%rsp)
vmovaps -0x70(%rsp), %xmm2
vmovaps 0x10(%rsp), %xmm0
vmovaps %xmm2, 0x2d0(%rsp)
vmovaps %xmm0, 0x2c0(%rsp)
vbroadcastss 0x7bac46(%rip), %xmm0 # 0x1e0be0c
vmovaps %xmm0, 0x2b0(%rsp)
vmovaps 0x2d0(%rsp), %xmm3
vmovaps 0x2c0(%rsp), %xmm2
vmovaps 0x2b0(%rsp), %xmm0
vmovaps %xmm3, 0x510(%rsp)
vmovaps %xmm2, 0x500(%rsp)
vmovaps %xmm0, 0x4f0(%rsp)
vmovaps 0x510(%rsp), %xmm2
vmovaps 0x500(%rsp), %xmm0
vmovaps 0x4f0(%rsp), %xmm3
vfmadd213ps %xmm3, %xmm2, %xmm0 # xmm0 = (xmm2 * xmm0) + xmm3
vmovaps %xmm0, -0x70(%rsp)
vmovaps -0x70(%rsp), %xmm2
vmovaps 0x10(%rsp), %xmm0
vmovaps %xmm2, 0x2a0(%rsp)
vmovaps %xmm0, 0x290(%rsp)
vbroadcastss 0x7babbe(%rip), %xmm0 # 0x1e0be10
vmovaps %xmm0, 0x280(%rsp)
vmovaps 0x2a0(%rsp), %xmm3
vmovaps 0x290(%rsp), %xmm2
vmovaps 0x280(%rsp), %xmm0
vmovaps %xmm3, 0x540(%rsp)
vmovaps %xmm2, 0x530(%rsp)
vmovaps %xmm0, 0x520(%rsp)
vmovaps 0x540(%rsp), %xmm2
vmovaps 0x530(%rsp), %xmm0
vmovaps 0x520(%rsp), %xmm3
vfmadd213ps %xmm3, %xmm2, %xmm0 # xmm0 = (xmm2 * xmm0) + xmm3
vmovaps %xmm0, -0x70(%rsp)
vmovaps -0x70(%rsp), %xmm2
vmovaps 0x10(%rsp), %xmm0
vmovaps %xmm2, 0x270(%rsp)
vmovaps %xmm0, 0x260(%rsp)
vbroadcastss 0x7bab36(%rip), %xmm0 # 0x1e0be14
vmovaps %xmm0, 0x250(%rsp)
vmovaps 0x270(%rsp), %xmm3
vmovaps 0x260(%rsp), %xmm2
vmovaps 0x250(%rsp), %xmm0
vmovaps %xmm3, 0x570(%rsp)
vmovaps %xmm2, 0x560(%rsp)
vmovaps %xmm0, 0x550(%rsp)
vmovaps 0x570(%rsp), %xmm2
vmovaps 0x560(%rsp), %xmm0
vmovaps 0x550(%rsp), %xmm3
vfmadd213ps %xmm3, %xmm2, %xmm0 # xmm0 = (xmm2 * xmm0) + xmm3
vmovaps %xmm0, -0x70(%rsp)
vmovaps -0x70(%rsp), %xmm2
vmovaps 0x10(%rsp), %xmm0
vmovaps %xmm2, 0xa0(%rsp)
vmovaps %xmm0, 0x90(%rsp)
vmovaps 0xa0(%rsp), %xmm0
vmovaps 0x90(%rsp), %xmm2
vmulps %xmm2, %xmm0, %xmm0
vmovaps %xmm0, -0x70(%rsp)
vmovaps -0x70(%rsp), %xmm2
vmovaps -0x60(%rsp), %xmm0
vmovaps %xmm2, 0x80(%rsp)
vmovaps %xmm0, 0x70(%rsp)
vmovaps 0x80(%rsp), %xmm0
vmovaps 0x70(%rsp), %xmm2
vmulps %xmm2, %xmm0, %xmm0
vmovaps %xmm0, -0x70(%rsp)
vmovaps -0x30(%rsp), %xmm2
vmovaps -0x70(%rsp), %xmm0
vmovaps %xmm2, 0x240(%rsp)
vbroadcastss 0x7baa01(%rip), %xmm2 # 0x1e0bdd0
vmovaps %xmm2, 0x230(%rsp)
vmovaps %xmm0, 0x220(%rsp)
vmovaps 0x240(%rsp), %xmm3
vmovaps 0x230(%rsp), %xmm2
vmovaps 0x220(%rsp), %xmm0
vmovaps %xmm3, 0x5a0(%rsp)
vmovaps %xmm2, 0x590(%rsp)
vmovaps %xmm0, 0x580(%rsp)
vmovaps 0x5a0(%rsp), %xmm2
vmovaps 0x590(%rsp), %xmm0
vmovaps 0x580(%rsp), %xmm3
vfmadd213ps %xmm3, %xmm2, %xmm0 # xmm0 = (xmm2 * xmm0) + xmm3
vmovaps %xmm0, -0x70(%rsp)
vmovaps -0x60(%rsp), %xmm2
vmovaps -0x70(%rsp), %xmm0
vmovaps %xmm2, 0x1e0(%rsp)
vmovaps %xmm1, 0x1d0(%rsp)
vmovaps %xmm0, 0x1c0(%rsp)
vmovaps 0x1e0(%rsp), %xmm2
vmovaps 0x1d0(%rsp), %xmm1
vmovaps 0x1c0(%rsp), %xmm0
vmovaps %xmm2, 0x3f0(%rsp)
vmovaps %xmm1, 0x3e0(%rsp)
vmovaps %xmm0, 0x3d0(%rsp)
vmovaps 0x3f0(%rsp), %xmm1
vmovaps 0x3e0(%rsp), %xmm0
vmovaps 0x3d0(%rsp), %xmm2
vfnmadd213ps %xmm2, %xmm1, %xmm0 # xmm0 = -(xmm1 * xmm0) + xmm2
vmovaps %xmm0, -0x70(%rsp)
vmovaps 0x10(%rsp), %xmm1
vmovaps -0x70(%rsp), %xmm0
vmovaps %xmm1, 0x100(%rsp)
vmovaps %xmm0, 0xf0(%rsp)
vmovaps 0x100(%rsp), %xmm0
vmovaps 0xf0(%rsp), %xmm1
vaddps %xmm1, %xmm0, %xmm0
vmovaps %xmm0, 0x10(%rsp)
vmovaps -0x30(%rsp), %xmm1
vmovaps 0x10(%rsp), %xmm0
vmovaps %xmm1, 0x210(%rsp)
vbroadcastss 0x7ba8b4(%rip), %xmm1 # 0x1e0bdcc
vmovaps %xmm1, 0x200(%rsp)
vmovaps %xmm0, 0x1f0(%rsp)
vmovaps 0x210(%rsp), %xmm2
vmovaps 0x200(%rsp), %xmm1
vmovaps 0x1f0(%rsp), %xmm0
vmovaps %xmm2, 0x5d0(%rsp)
vmovaps %xmm1, 0x5c0(%rsp)
vmovaps %xmm0, 0x5b0(%rsp)
vmovaps 0x5d0(%rsp), %xmm1
vmovaps 0x5c0(%rsp), %xmm0
vmovaps 0x5b0(%rsp), %xmm2
vfmadd213ps %xmm2, %xmm1, %xmm0 # xmm0 = (xmm1 * xmm0) + xmm2
vmovaps %xmm0, 0x10(%rsp)
vmovaps 0x10(%rsp), %xmm1
vmovaps -0x20(%rsp), %xmm0
vmovaps %xmm1, 0x640(%rsp)
vmovaps %xmm0, 0x630(%rsp)
vmovaps 0x640(%rsp), %xmm0
vmovaps 0x630(%rsp), %xmm1
vpor %xmm1, %xmm0, %xmm0
vmovaps %xmm0, 0x10(%rsp)
vmovaps 0x10(%rsp), %xmm0
addq $0x6b8, %rsp # imm = 0x6B8
retq
nop
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,288 |
ncnn::UnaryOp_x86_avx512_functor::unary_op_log::func(float const&) const
|
float func(const float& x) const
{
return (float)log(x);
}
|
subq $0x18, %rsp
movq %rdi, 0x10(%rsp)
movq %rsi, 0x8(%rsp)
movq 0x8(%rsp), %rax
vmovss (%rax), %xmm0
callq 0x100cd0
addq $0x18, %rsp
retq
nopw %cs:(%rax,%rax)
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,289 |
ncnn::UnaryOp_x86_avx512_functor::unary_op_sin::func_pack16(float vector[16] const&) const
|
__m512 func_pack16(const __m512& x) const
{
return sin512_ps(x);
}
|
pushq %rbp
movq %rsp, %rbp
andq $-0x40, %rsp
subq $0x15c0, %rsp # imm = 0x15C0
movq %rdi, 0x38(%rsp)
movq %rsi, 0x30(%rsp)
movq 0x30(%rsp), %rax
vmovaps (%rax), %zmm0
vmovaps %zmm0, 0x300(%rsp)
vmovaps 0x300(%rsp), %zmm0
vmovaps %zmm0, 0x200(%rsp)
vmovaps 0x300(%rsp), %zmm0
vmovaps %zmm0, 0x480(%rsp)
vmovaps 0x480(%rsp), %zmm1
movl $0x7fffffff, 0x53c(%rsp) # imm = 0x7FFFFFFF
vpbroadcastd 0x53c(%rsp), %zmm0
vmovdqa64 %zmm0, 0x4c0(%rsp)
vmovdqa64 0x4c0(%rsp), %zmm0
vmovaps %zmm1, 0x440(%rsp)
vmovdqa64 %zmm0, 0x400(%rsp)
vmovdqa64 0x440(%rsp), %zmm0
vmovdqa64 0x400(%rsp), %zmm1
vpandq %zmm1, %zmm0, %zmm0
vmovdqa64 %zmm0, 0x3c0(%rsp)
vmovdqa64 0x3c0(%rsp), %zmm0
vmovdqa64 %zmm0, 0x300(%rsp)
vmovaps 0x200(%rsp), %zmm1
vmovaps 0x7cc3f6(%rip), %zmm0 # 0x1e1dac0
vmovaps %zmm1, 0x11c0(%rsp)
vmovaps %zmm0, 0x1180(%rsp)
vmovdqa64 0x11c0(%rsp), %zmm0
vmovdqa64 0x1180(%rsp), %zmm1
vpandd %zmm1, %zmm0, %zmm0
vmovdqa64 %zmm0, 0x200(%rsp)
vmovaps 0x300(%rsp), %zmm1
vmovaps 0x7cc3f6(%rip), %zmm0 # 0x1e1db00
vmovaps %zmm1, 0x780(%rsp)
vmovaps %zmm0, 0x740(%rsp)
vmovaps 0x780(%rsp), %zmm0
vmovaps 0x740(%rsp), %zmm1
vmulps %zmm1, %zmm0, %zmm0
vmovaps %zmm0, 0x1c0(%rsp)
vmovaps 0x1c0(%rsp), %zmm0
vmovaps %zmm0, 0xf80(%rsp)
vpxor %xmm1, %xmm1, %xmm1
vcvttps2dq 0xf80(%rsp), %zmm0
vmovdqa64 %zmm1, 0x10c0(%rsp)
vmovdqa64 %zmm0, 0x140(%rsp)
vmovdqa64 0x140(%rsp), %zmm1
vmovdqa64 0x7cc3ca(%rip), %zmm0 # 0x1e1db40
vmovdqa64 %zmm1, 0x1000(%rsp)
vmovdqa64 %zmm0, 0xfc0(%rsp)
vmovdqa64 0x1000(%rsp), %zmm0
vmovdqa64 0xfc0(%rsp), %zmm1
vpaddd %zmm1, %zmm0, %zmm0
vmovdqa64 %zmm0, 0x140(%rsp)
vmovdqa64 0x140(%rsp), %zmm1
vmovdqa64 0x7cc3ca(%rip), %zmm0 # 0x1e1db80
vmovdqa64 %zmm1, 0x13c0(%rsp)
vmovdqa64 %zmm0, 0x1380(%rsp)
vmovdqa64 0x13c0(%rsp), %zmm0
vmovdqa64 0x1380(%rsp), %zmm1
vpandq %zmm1, %zmm0, %zmm0
vmovdqa64 %zmm0, 0x140(%rsp)
vmovdqa64 0x140(%rsp), %zmm0
vmovdqa64 %zmm0, 0x1200(%rsp)
vcvtdq2ps 0x1200(%rsp), %zmm0
vmovaps %zmm0, 0x1c0(%rsp)
vmovdqa64 0x140(%rsp), %zmm1
vmovdqa64 0x7cc3aa(%rip), %zmm0 # 0x1e1dbc0
vmovdqa64 %zmm1, 0x1340(%rsp)
vmovdqa64 %zmm0, 0x1300(%rsp)
vmovdqa64 0x1340(%rsp), %zmm0
vmovdqa64 0x1300(%rsp), %zmm1
vpandq %zmm1, %zmm0, %zmm0
vmovdqa64 %zmm0, 0x180(%rsp)
vmovdqa64 0x180(%rsp), %zmm0
vmovdqa64 %zmm0, 0x1080(%rsp)
movl $0x1d, 0x107c(%rsp)
vmovdqa64 0x1080(%rsp), %zmm0
vmovd 0x107c(%rsp), %xmm1
vpslld %xmm1, %zmm0, %zmm0
vmovdqa64 %zmm0, 0x180(%rsp)
vmovdqa64 0x140(%rsp), %zmm1
vmovdqa64 0x7cc370(%rip), %zmm0 # 0x1e1dc00
vmovdqa64 %zmm1, 0x12c0(%rsp)
vmovdqa64 %zmm0, 0x1280(%rsp)
vmovdqa64 0x12c0(%rsp), %zmm0
vmovdqa64 0x1280(%rsp), %zmm1
vpandq %zmm1, %zmm0, %zmm0
vmovdqa64 %zmm0, 0x140(%rsp)
vmovdqa64 0x140(%rsp), %zmm0
vptestnmd %zmm0, %zmm0, %k0
kmovw %k0, 0x127e(%rsp)
kmovw 0x127e(%rsp), %k0
vpmovm2d %k0, %zmm0
vmovdqa64 %zmm0, 0x140(%rsp)
vmovdqa64 0x180(%rsp), %zmm0
vmovdqa64 %zmm0, 0x380(%rsp)
vmovdqa64 0x380(%rsp), %zmm0
vmovdqa64 %zmm0, 0x100(%rsp)
vmovdqa64 0x140(%rsp), %zmm0
vmovdqa64 %zmm0, 0x340(%rsp)
vmovdqa64 0x340(%rsp), %zmm0
vmovdqa64 %zmm0, 0xc0(%rsp)
vmovaps 0x200(%rsp), %zmm1
vmovaps 0x100(%rsp), %zmm0
vmovaps %zmm1, 0x14c0(%rsp)
vmovaps %zmm0, 0x1480(%rsp)
vmovdqa64 0x14c0(%rsp), %zmm0
vmovdqa64 0x1480(%rsp), %zmm1
vpxord %zmm1, %zmm0, %zmm0
vmovdqa64 %zmm0, 0x200(%rsp)
vmovaps 0x7cc30c(%rip), %zmm0 # 0x1e1dc80
vmovaps %zmm0, 0x2c0(%rsp)
vmovaps 0x7cc33a(%rip), %zmm0 # 0x1e1dcc0
vmovaps %zmm0, 0x280(%rsp)
vmovaps 0x7cc368(%rip), %zmm0 # 0x1e1dd00
vmovaps %zmm0, 0x240(%rsp)
vmovaps 0x1c0(%rsp), %zmm2
vmovaps 0x2c0(%rsp), %zmm1
vmovaps 0x300(%rsp), %zmm0
vmovaps %zmm2, 0xd80(%rsp)
vmovaps %zmm1, 0xd40(%rsp)
vmovaps %zmm0, 0xd00(%rsp)
vmovaps 0xd80(%rsp), %zmm1
vmovaps 0xd40(%rsp), %zmm0
vmovaps 0xd00(%rsp), %zmm2
vfmadd213ps %zmm2, %zmm1, %zmm0 # zmm0 = (zmm1 * zmm0) + zmm2
vmovaps %zmm0, 0x300(%rsp)
vmovaps 0x1c0(%rsp), %zmm2
vmovaps 0x280(%rsp), %zmm1
vmovaps 0x300(%rsp), %zmm0
vmovaps %zmm2, 0xcc0(%rsp)
vmovaps %zmm1, 0xc80(%rsp)
vmovaps %zmm0, 0xc40(%rsp)
vmovaps 0xcc0(%rsp), %zmm1
vmovaps 0xc80(%rsp), %zmm0
vmovaps 0xc40(%rsp), %zmm2
vfmadd213ps %zmm2, %zmm1, %zmm0 # zmm0 = (zmm1 * zmm0) + zmm2
vmovaps %zmm0, 0x300(%rsp)
vmovaps 0x1c0(%rsp), %zmm2
vmovaps 0x240(%rsp), %zmm1
vmovaps 0x300(%rsp), %zmm0
vmovaps %zmm2, 0xc00(%rsp)
vmovaps %zmm1, 0xbc0(%rsp)
vmovaps %zmm0, 0xb80(%rsp)
vmovaps 0xc00(%rsp), %zmm1
vmovaps 0xbc0(%rsp), %zmm0
vmovaps 0xb80(%rsp), %zmm2
vfmadd213ps %zmm2, %zmm1, %zmm0 # zmm0 = (zmm1 * zmm0) + zmm2
vmovaps %zmm0, 0x300(%rsp)
vmovaps 0x7cc294(%rip), %zmm0 # 0x1e1dd40
vmovaps %zmm0, 0x1c0(%rsp)
vmovaps 0x300(%rsp), %zmm0
vmovaps %zmm0, 0x700(%rsp)
vmovaps %zmm0, 0x6c0(%rsp)
vmovaps 0x700(%rsp), %zmm0
vmovaps 0x6c0(%rsp), %zmm1
vmulps %zmm1, %zmm0, %zmm0
vmovaps %zmm0, 0x80(%rsp)
vmovaps 0x1c0(%rsp), %zmm2
vmovaps 0x80(%rsp), %zmm1
vmovaps 0x7cc27c(%rip), %zmm0 # 0x1e1dd80
vmovaps %zmm2, 0xb40(%rsp)
vmovaps %zmm1, 0xb00(%rsp)
vmovaps %zmm0, 0xac0(%rsp)
vmovaps 0xb40(%rsp), %zmm1
vmovaps 0xb00(%rsp), %zmm0
vmovaps 0xac0(%rsp), %zmm2
vfmadd213ps %zmm2, %zmm1, %zmm0 # zmm0 = (zmm1 * zmm0) + zmm2
vmovaps %zmm0, 0x1c0(%rsp)
vmovaps 0x1c0(%rsp), %zmm2
vmovaps 0x80(%rsp), %zmm1
vmovaps 0x7cc264(%rip), %zmm0 # 0x1e1ddc0
vmovaps %zmm2, 0xa80(%rsp)
vmovaps %zmm1, 0xa40(%rsp)
vmovaps %zmm0, 0xa00(%rsp)
vmovaps 0xa80(%rsp), %zmm1
vmovaps 0xa40(%rsp), %zmm0
vmovaps 0xa00(%rsp), %zmm2
vfmadd213ps %zmm2, %zmm1, %zmm0 # zmm0 = (zmm1 * zmm0) + zmm2
vmovaps %zmm0, 0x1c0(%rsp)
vmovaps 0x1c0(%rsp), %zmm1
vmovaps 0x80(%rsp), %zmm0
vmovaps %zmm1, 0x680(%rsp)
vmovaps %zmm0, 0x640(%rsp)
vmovaps 0x680(%rsp), %zmm0
vmovaps 0x640(%rsp), %zmm1
vmulps %zmm1, %zmm0, %zmm0
vmovaps %zmm0, 0x1c0(%rsp)
vmovaps 0x1c0(%rsp), %zmm1
vmovaps 0x80(%rsp), %zmm0
vmovaps %zmm1, 0x600(%rsp)
vmovaps %zmm0, 0x5c0(%rsp)
vmovaps 0x600(%rsp), %zmm0
vmovaps 0x5c0(%rsp), %zmm1
vmulps %zmm1, %zmm0, %zmm0
vmovaps %zmm0, 0x1c0(%rsp)
vmovaps 0x80(%rsp), %zmm2
vmovaps 0x7cb318(%rip), %zmm1 # 0x1e1cf40
vmovaps 0x1c0(%rsp), %zmm0
vmovaps %zmm2, 0xe40(%rsp)
vmovaps %zmm1, 0xe00(%rsp)
vmovaps %zmm0, 0xdc0(%rsp)
vmovaps 0xe40(%rsp), %zmm1
vmovaps 0xe00(%rsp), %zmm0
vmovaps 0xdc0(%rsp), %zmm2
vfnmadd213ps %zmm2, %zmm1, %zmm0 # zmm0 = -(zmm1 * zmm0) + zmm2
vmovaps %zmm0, 0x1c0(%rsp)
vmovaps 0x1c0(%rsp), %zmm1
vmovaps 0x7cb1c0(%rip), %zmm0 # 0x1e1ce40
vmovaps %zmm1, 0xf40(%rsp)
vmovaps %zmm0, 0xf00(%rsp)
vmovaps 0xf40(%rsp), %zmm0
vmovaps 0xf00(%rsp), %zmm1
vaddps %zmm1, %zmm0, %zmm0
vmovaps %zmm0, 0x1c0(%rsp)
vmovaps 0x7cc148(%rip), %zmm0 # 0x1e1de00
vmovaps %zmm0, 0x40(%rsp)
vmovaps 0x40(%rsp), %zmm2
vmovaps 0x80(%rsp), %zmm1
vmovaps 0x7cc166(%rip), %zmm0 # 0x1e1de40
vmovaps %zmm2, 0x9c0(%rsp)
vmovaps %zmm1, 0x980(%rsp)
vmovaps %zmm0, 0x940(%rsp)
vmovaps 0x9c0(%rsp), %zmm1
vmovaps 0x980(%rsp), %zmm0
vmovaps 0x940(%rsp), %zmm2
vfmadd213ps %zmm2, %zmm1, %zmm0 # zmm0 = (zmm1 * zmm0) + zmm2
vmovaps %zmm0, 0x40(%rsp)
vmovaps 0x40(%rsp), %zmm2
vmovaps 0x80(%rsp), %zmm1
vmovaps 0x7cc14e(%rip), %zmm0 # 0x1e1de80
vmovaps %zmm2, 0x900(%rsp)
vmovaps %zmm1, 0x8c0(%rsp)
vmovaps %zmm0, 0x880(%rsp)
vmovaps 0x900(%rsp), %zmm1
vmovaps 0x8c0(%rsp), %zmm0
vmovaps 0x880(%rsp), %zmm2
vfmadd213ps %zmm2, %zmm1, %zmm0 # zmm0 = (zmm1 * zmm0) + zmm2
vmovaps %zmm0, 0x40(%rsp)
vmovaps 0x40(%rsp), %zmm1
vmovaps 0x80(%rsp), %zmm0
vmovaps %zmm1, 0x580(%rsp)
vmovaps %zmm0, 0x540(%rsp)
vmovaps 0x580(%rsp), %zmm0
vmovaps 0x540(%rsp), %zmm1
vmulps %zmm1, %zmm0, %zmm0
vmovaps %zmm0, 0x40(%rsp)
vmovaps 0x40(%rsp), %zmm1
vmovaps 0x300(%rsp), %zmm0
vmovaps %zmm1, 0x840(%rsp)
vmovaps %zmm0, 0x800(%rsp)
vmovaps %zmm0, 0x7c0(%rsp)
vmovaps 0x840(%rsp), %zmm1
vmovaps 0x800(%rsp), %zmm0
vmovaps 0x7c0(%rsp), %zmm2
vfmadd213ps %zmm2, %zmm1, %zmm0 # zmm0 = (zmm1 * zmm0) + zmm2
vmovaps %zmm0, 0x40(%rsp)
vmovaps 0xc0(%rsp), %zmm0
vmovaps %zmm0, 0x240(%rsp)
vmovaps 0x240(%rsp), %zmm1
vmovaps 0x40(%rsp), %zmm0
vmovaps %zmm1, 0x1140(%rsp)
vmovaps %zmm0, 0x1100(%rsp)
vmovdqa64 0x1140(%rsp), %zmm0
vmovdqa64 0x1100(%rsp), %zmm1
vpandd %zmm1, %zmm0, %zmm0
vmovdqa64 %zmm0, 0x40(%rsp)
vmovaps 0x240(%rsp), %zmm1
vmovaps 0x1c0(%rsp), %zmm0
vmovaps %zmm1, 0x1540(%rsp)
vmovaps %zmm0, 0x1500(%rsp)
vmovdqa64 0x1540(%rsp), %zmm0
vpternlogq $0xf, %zmm0, %zmm0, %zmm0
vmovaps 0x1500(%rsp), %zmm1
vpandd %zmm1, %zmm0, %zmm0
vmovaps %zmm0, 0x1c0(%rsp)
vmovaps 0x1c0(%rsp), %zmm1
vmovaps 0x40(%rsp), %zmm0
vmovaps %zmm1, 0xec0(%rsp)
vmovaps %zmm0, 0xe80(%rsp)
vmovaps 0xec0(%rsp), %zmm0
vaddps 0xe80(%rsp), %zmm0, %zmm0
vmovaps %zmm0, 0x1c0(%rsp)
vmovaps 0x1c0(%rsp), %zmm1
vmovaps 0x200(%rsp), %zmm0
vmovaps %zmm1, 0x1440(%rsp)
vmovaps %zmm0, 0x1400(%rsp)
vmovaps 0x1440(%rsp), %zmm0
vmovaps 0x1400(%rsp), %zmm1
vpxord %zmm1, %zmm0, %zmm0
vmovaps %zmm0, 0x1c0(%rsp)
vmovaps 0x1c0(%rsp), %zmm0
movq %rbp, %rsp
popq %rbp
retq
nopw %cs:(%rax,%rax)
nopl (%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,290 |
ncnn::UnaryOp_x86_avx512_functor::unary_op_sin::func_pack8(float vector[8] const&) const
|
__m256 func_pack8(const __m256& x) const
{
return sin256_ps(x);
}
|
pushq %rbp
movq %rsp, %rbp
andq $-0x20, %rsp
subq $0xe60, %rsp # imm = 0xE60
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
movq 0x10(%rsp), %rax
vmovaps (%rax), %ymm0
vmovaps %ymm0, 0x180(%rsp)
vpxor %xmm0, %xmm0, %xmm0
vmovaps %ymm0, 0x1a0(%rsp)
vmovaps 0x1a0(%rsp), %ymm0
vmovaps %ymm0, 0x140(%rsp)
vmovaps 0x180(%rsp), %ymm0
vmovaps %ymm0, 0x100(%rsp)
vmovaps 0x180(%rsp), %ymm0
vmovaps %ymm0, 0x6a0(%rsp)
vbroadcastss 0x7cae19(%rip), %ymm0 # 0x1e1cdb0
vmovaps %ymm0, 0x680(%rsp)
vmovdqa 0x6a0(%rsp), %ymm0
vmovdqa 0x680(%rsp), %ymm1
vpand %ymm1, %ymm0, %ymm0
vmovdqa %ymm0, 0x180(%rsp)
vmovaps 0x100(%rsp), %ymm0
vmovaps %ymm0, 0x660(%rsp)
vbroadcastss 0x7bbede(%rip), %ymm0 # 0x1e0deb8
vmovaps %ymm0, 0x640(%rsp)
vmovdqa 0x660(%rsp), %ymm0
vmovdqa 0x640(%rsp), %ymm1
vpand %ymm1, %ymm0, %ymm0
vmovdqa %ymm0, 0x100(%rsp)
vmovaps 0x180(%rsp), %ymm0
vmovaps %ymm0, 0x2e0(%rsp)
vbroadcastss 0x7cad9b(%rip), %ymm0 # 0x1e1cdb8
vmovaps %ymm0, 0x2c0(%rsp)
vmovaps 0x2e0(%rsp), %ymm0
vmovaps 0x2c0(%rsp), %ymm1
vmulps %ymm1, %ymm0, %ymm0
vmovaps %ymm0, 0xe0(%rsp)
vmovaps 0xe0(%rsp), %ymm0
vmovaps %ymm0, 0x7a0(%rsp)
vcvttps2dq 0x7a0(%rsp), %ymm0
vmovdqa %ymm0, 0xa0(%rsp)
vmovdqa 0xa0(%rsp), %ymm0
vmovdqa %ymm0, 0x7e0(%rsp)
vpbroadcastq 0x7cad5c(%rip), %ymm0 # 0x1e1cde0
vmovdqa %ymm0, 0x7c0(%rsp)
vmovdqa 0x7e0(%rsp), %ymm1
vmovdqa 0x7c0(%rsp), %ymm0
vmovdqa %ymm1, 0xc00(%rsp)
vmovdqa %ymm0, 0xbe0(%rsp)
vmovdqa 0xc00(%rsp), %ymm0
vmovdqa 0xbe0(%rsp), %ymm1
vpaddd %ymm1, %ymm0, %ymm0
vmovdqa %ymm0, 0xa0(%rsp)
vmovdqa 0xa0(%rsp), %ymm0
vmovdqa %ymm0, 0xd20(%rsp)
vpbroadcastq 0x7cacfd(%rip), %ymm0 # 0x1e1cde8
vmovdqa %ymm0, 0xd00(%rsp)
vmovdqa 0xd20(%rsp), %ymm0
vmovdqa 0xd00(%rsp), %ymm1
vpand %ymm1, %ymm0, %ymm0
vmovdqa %ymm0, 0xa0(%rsp)
vmovdqa 0xa0(%rsp), %ymm0
vmovdqa %ymm0, 0xc60(%rsp)
vcvtdq2ps 0xc60(%rsp), %ymm0
vmovaps %ymm0, 0xe0(%rsp)
vmovdqa 0xa0(%rsp), %ymm0
vmovdqa %ymm0, 0xce0(%rsp)
vpbroadcastq 0x7cac9e(%rip), %ymm0 # 0x1e1cdf0
vmovdqa %ymm0, 0xcc0(%rsp)
vmovdqa 0xce0(%rsp), %ymm0
vmovdqa 0xcc0(%rsp), %ymm1
vpand %ymm1, %ymm0, %ymm0
vmovdqa %ymm0, 0xc0(%rsp)
vmovdqa 0xc0(%rsp), %ymm0
vmovdqa %ymm0, 0x820(%rsp)
movl $0x1d, 0x81c(%rsp)
vmovdqa 0x820(%rsp), %ymm0
movl 0x81c(%rsp), %eax
vmovdqa %ymm0, 0xc40(%rsp)
movl %eax, 0xc3c(%rsp)
vmovdqa 0xc40(%rsp), %ymm0
vmovd 0xc3c(%rsp), %xmm1
vpslld %xmm1, %ymm0, %ymm0
vmovdqa %ymm0, 0xc0(%rsp)
vmovdqa 0xa0(%rsp), %ymm0
vmovdqa %ymm0, 0xca0(%rsp)
vpbroadcastq 0x7cac07(%rip), %ymm0 # 0x1e1cdf8
vmovdqa %ymm0, 0xc80(%rsp)
vmovdqa 0xca0(%rsp), %ymm0
vmovdqa 0xc80(%rsp), %ymm1
vpand %ymm1, %ymm0, %ymm0
vmovdqa %ymm0, 0xa0(%rsp)
vmovdqa 0xa0(%rsp), %ymm0
vmovdqa %ymm0, 0xd60(%rsp)
vpxor %xmm0, %xmm0, %xmm0
vmovdqa %ymm0, 0xd40(%rsp)
vmovdqa 0xd60(%rsp), %ymm0
vmovdqa 0xd40(%rsp), %ymm1
vpcmpeqd %ymm1, %ymm0, %ymm0
vmovdqa %ymm0, 0xa0(%rsp)
vmovdqa 0xc0(%rsp), %ymm0
vmovdqa %ymm0, 0x860(%rsp)
vmovdqa 0x860(%rsp), %ymm0
vmovdqa %ymm0, 0x80(%rsp)
vmovdqa 0xa0(%rsp), %ymm0
vmovdqa %ymm0, 0x840(%rsp)
vmovdqa 0x840(%rsp), %ymm0
vmovdqa %ymm0, 0x60(%rsp)
vmovaps 0x100(%rsp), %ymm1
vmovaps 0x80(%rsp), %ymm0
vmovaps %ymm1, 0xde0(%rsp)
vmovaps %ymm0, 0xdc0(%rsp)
vmovdqa 0xde0(%rsp), %ymm0
vmovdqa 0xdc0(%rsp), %ymm1
vpxor %ymm1, %ymm0, %ymm0
vmovdqa %ymm0, 0x100(%rsp)
vbroadcastss 0x7caad4(%rip), %ymm0 # 0x1e1cdbc
vmovaps %ymm0, 0x160(%rsp)
vbroadcastss 0x7caac6(%rip), %ymm0 # 0x1e1cdc0
vmovaps %ymm0, 0x140(%rsp)
vbroadcastss 0x7caab8(%rip), %ymm0 # 0x1e1cdc4
vmovaps %ymm0, 0x120(%rsp)
vmovaps 0xe0(%rsp), %ymm2
vmovaps 0x160(%rsp), %ymm1
vmovaps 0x180(%rsp), %ymm0
vmovaps %ymm2, 0x5e0(%rsp)
vmovaps %ymm1, 0x5c0(%rsp)
vmovaps %ymm0, 0x5a0(%rsp)
vmovaps 0x5e0(%rsp), %ymm2
vmovaps 0x5c0(%rsp), %ymm1
vmovaps 0x5a0(%rsp), %ymm0
vmovaps %ymm2, 0x8c0(%rsp)
vmovaps %ymm1, 0x8a0(%rsp)
vmovaps %ymm0, 0x880(%rsp)
vmovaps 0x8c0(%rsp), %ymm1
vmovaps 0x8a0(%rsp), %ymm0
vmovaps 0x880(%rsp), %ymm2
vfmadd213ps %ymm2, %ymm1, %ymm0 # ymm0 = (ymm1 * ymm0) + ymm2
vmovaps %ymm0, 0x180(%rsp)
vmovaps 0xe0(%rsp), %ymm2
vmovaps 0x140(%rsp), %ymm1
vmovaps 0x180(%rsp), %ymm0
vmovaps %ymm2, 0x580(%rsp)
vmovaps %ymm1, 0x560(%rsp)
vmovaps %ymm0, 0x540(%rsp)
vmovaps 0x580(%rsp), %ymm2
vmovaps 0x560(%rsp), %ymm1
vmovaps 0x540(%rsp), %ymm0
vmovaps %ymm2, 0x920(%rsp)
vmovaps %ymm1, 0x900(%rsp)
vmovaps %ymm0, 0x8e0(%rsp)
vmovaps 0x920(%rsp), %ymm1
vmovaps 0x900(%rsp), %ymm0
vmovaps 0x8e0(%rsp), %ymm2
vfmadd213ps %ymm2, %ymm1, %ymm0 # ymm0 = (ymm1 * ymm0) + ymm2
vmovaps %ymm0, 0x180(%rsp)
vmovaps 0xe0(%rsp), %ymm2
vmovaps 0x120(%rsp), %ymm1
vmovaps 0x180(%rsp), %ymm0
vmovaps %ymm2, 0x520(%rsp)
vmovaps %ymm1, 0x500(%rsp)
vmovaps %ymm0, 0x4e0(%rsp)
vmovaps 0x520(%rsp), %ymm2
vmovaps 0x500(%rsp), %ymm1
vmovaps 0x4e0(%rsp), %ymm0
vmovaps %ymm2, 0x980(%rsp)
vmovaps %ymm1, 0x960(%rsp)
vmovaps %ymm0, 0x940(%rsp)
vmovaps 0x980(%rsp), %ymm1
vmovaps 0x960(%rsp), %ymm0
vmovaps 0x940(%rsp), %ymm2
vfmadd213ps %ymm2, %ymm1, %ymm0 # ymm0 = (ymm1 * ymm0) + ymm2
vmovaps %ymm0, 0x180(%rsp)
vbroadcastss 0x7ca8eb(%rip), %ymm0 # 0x1e1cdc8
vmovaps %ymm0, 0xe0(%rsp)
vmovaps 0x180(%rsp), %ymm0
vmovaps %ymm0, 0x2a0(%rsp)
vmovaps %ymm0, 0x280(%rsp)
vmovaps 0x2a0(%rsp), %ymm0
vmovaps 0x280(%rsp), %ymm1
vmulps %ymm1, %ymm0, %ymm0
vmovaps %ymm0, 0x40(%rsp)
vmovaps 0xe0(%rsp), %ymm1
vmovaps 0x40(%rsp), %ymm0
vmovaps %ymm1, 0x4c0(%rsp)
vmovaps %ymm0, 0x4a0(%rsp)
vbroadcastss 0x7ca885(%rip), %ymm0 # 0x1e1cdcc
vmovaps %ymm0, 0x480(%rsp)
vmovaps 0x4c0(%rsp), %ymm2
vmovaps 0x4a0(%rsp), %ymm1
vmovaps 0x480(%rsp), %ymm0
vmovaps %ymm2, 0x9e0(%rsp)
vmovaps %ymm1, 0x9c0(%rsp)
vmovaps %ymm0, 0x9a0(%rsp)
vmovaps 0x9e0(%rsp), %ymm1
vmovaps 0x9c0(%rsp), %ymm0
vmovaps 0x9a0(%rsp), %ymm2
vfmadd213ps %ymm2, %ymm1, %ymm0 # ymm0 = (ymm1 * ymm0) + ymm2
vmovaps %ymm0, 0xe0(%rsp)
vmovaps 0xe0(%rsp), %ymm1
vmovaps 0x40(%rsp), %ymm0
vmovaps %ymm1, 0x460(%rsp)
vmovaps %ymm0, 0x440(%rsp)
vbroadcastss 0x7ca7f7(%rip), %ymm0 # 0x1e1cdd0
vmovaps %ymm0, 0x420(%rsp)
vmovaps 0x460(%rsp), %ymm2
vmovaps 0x440(%rsp), %ymm1
vmovaps 0x420(%rsp), %ymm0
vmovaps %ymm2, 0xa40(%rsp)
vmovaps %ymm1, 0xa20(%rsp)
vmovaps %ymm0, 0xa00(%rsp)
vmovaps 0xa40(%rsp), %ymm1
vmovaps 0xa20(%rsp), %ymm0
vmovaps 0xa00(%rsp), %ymm2
vfmadd213ps %ymm2, %ymm1, %ymm0 # ymm0 = (ymm1 * ymm0) + ymm2
vmovaps %ymm0, 0xe0(%rsp)
vmovaps 0xe0(%rsp), %ymm1
vmovaps 0x40(%rsp), %ymm0
vmovaps %ymm1, 0x260(%rsp)
vmovaps %ymm0, 0x240(%rsp)
vmovaps 0x260(%rsp), %ymm0
vmovaps 0x240(%rsp), %ymm1
vmulps %ymm1, %ymm0, %ymm0
vmovaps %ymm0, 0xe0(%rsp)
vmovaps 0xe0(%rsp), %ymm1
vmovaps 0x40(%rsp), %ymm0
vmovaps %ymm1, 0x220(%rsp)
vmovaps %ymm0, 0x200(%rsp)
vmovaps 0x220(%rsp), %ymm0
vmovaps 0x200(%rsp), %ymm1
vmulps %ymm1, %ymm0, %ymm0
vmovaps %ymm0, 0xe0(%rsp)
vmovaps 0x40(%rsp), %ymm1
vmovaps 0xe0(%rsp), %ymm0
vmovaps %ymm1, 0x700(%rsp)
vbroadcastss 0x7b693a(%rip), %ymm1 # 0x1e0901c
vmovaps %ymm1, 0x6e0(%rsp)
vmovaps %ymm0, 0x6c0(%rsp)
vmovaps 0x700(%rsp), %ymm2
vmovaps 0x6e0(%rsp), %ymm1
vmovaps 0x6c0(%rsp), %ymm0
vmovaps %ymm2, 0xbc0(%rsp)
vmovaps %ymm1, 0xba0(%rsp)
vmovaps %ymm0, 0xb80(%rsp)
vmovaps 0xbc0(%rsp), %ymm1
vmovaps 0xba0(%rsp), %ymm0
vmovaps 0xb80(%rsp), %ymm2
vfnmadd213ps %ymm2, %ymm1, %ymm0 # ymm0 = -(ymm1 * ymm0) + ymm2
vmovaps %ymm0, 0xe0(%rsp)
vmovaps 0xe0(%rsp), %ymm0
vmovaps %ymm0, 0x780(%rsp)
vbroadcastss 0x7b6896(%rip), %ymm0 # 0x1e09004
vmovaps %ymm0, 0x760(%rsp)
vmovaps 0x780(%rsp), %ymm0
vmovaps 0x760(%rsp), %ymm1
vaddps %ymm1, %ymm0, %ymm0
vmovaps %ymm0, 0xe0(%rsp)
vbroadcastss 0x7ca635(%rip), %ymm0 # 0x1e1cdd4
vmovaps %ymm0, 0x20(%rsp)
vmovaps 0x20(%rsp), %ymm1
vmovaps 0x40(%rsp), %ymm0
vmovaps %ymm1, 0x400(%rsp)
vmovaps %ymm0, 0x3e0(%rsp)
vbroadcastss 0x7ca60c(%rip), %ymm0 # 0x1e1cdd8
vmovaps %ymm0, 0x3c0(%rsp)
vmovaps 0x400(%rsp), %ymm2
vmovaps 0x3e0(%rsp), %ymm1
vmovaps 0x3c0(%rsp), %ymm0
vmovaps %ymm2, 0xaa0(%rsp)
vmovaps %ymm1, 0xa80(%rsp)
vmovaps %ymm0, 0xa60(%rsp)
vmovaps 0xaa0(%rsp), %ymm1
vmovaps 0xa80(%rsp), %ymm0
vmovaps 0xa60(%rsp), %ymm2
vfmadd213ps %ymm2, %ymm1, %ymm0 # ymm0 = (ymm1 * ymm0) + ymm2
vmovaps %ymm0, 0x20(%rsp)
vmovaps 0x20(%rsp), %ymm1
vmovaps 0x40(%rsp), %ymm0
vmovaps %ymm1, 0x3a0(%rsp)
vmovaps %ymm0, 0x380(%rsp)
vbroadcastss 0x7ca584(%rip), %ymm0 # 0x1e1cddc
vmovaps %ymm0, 0x360(%rsp)
vmovaps 0x3a0(%rsp), %ymm2
vmovaps 0x380(%rsp), %ymm1
vmovaps 0x360(%rsp), %ymm0
vmovaps %ymm2, 0xb00(%rsp)
vmovaps %ymm1, 0xae0(%rsp)
vmovaps %ymm0, 0xac0(%rsp)
vmovaps 0xb00(%rsp), %ymm1
vmovaps 0xae0(%rsp), %ymm0
vmovaps 0xac0(%rsp), %ymm2
vfmadd213ps %ymm2, %ymm1, %ymm0 # ymm0 = (ymm1 * ymm0) + ymm2
vmovaps %ymm0, 0x20(%rsp)
vmovaps 0x20(%rsp), %ymm1
vmovaps 0x40(%rsp), %ymm0
vmovaps %ymm1, 0x1e0(%rsp)
vmovaps %ymm0, 0x1c0(%rsp)
vmovaps 0x1e0(%rsp), %ymm0
vmovaps 0x1c0(%rsp), %ymm1
vmulps %ymm1, %ymm0, %ymm0
vmovaps %ymm0, 0x20(%rsp)
vmovaps 0x20(%rsp), %ymm1
vmovaps 0x180(%rsp), %ymm0
vmovaps %ymm1, 0x340(%rsp)
vmovaps %ymm0, 0x320(%rsp)
vmovaps %ymm0, 0x300(%rsp)
vmovaps 0x340(%rsp), %ymm2
vmovaps 0x320(%rsp), %ymm1
vmovaps 0x300(%rsp), %ymm0
vmovaps %ymm2, 0xb60(%rsp)
vmovaps %ymm1, 0xb40(%rsp)
vmovaps %ymm0, 0xb20(%rsp)
vmovaps 0xb60(%rsp), %ymm1
vmovaps 0xb40(%rsp), %ymm0
vmovaps 0xb20(%rsp), %ymm2
vfmadd213ps %ymm2, %ymm1, %ymm0 # ymm0 = (ymm1 * ymm0) + ymm2
vmovaps %ymm0, 0x20(%rsp)
vmovaps 0x60(%rsp), %ymm0
vmovaps %ymm0, 0x120(%rsp)
vmovaps 0x120(%rsp), %ymm1
vmovaps 0x20(%rsp), %ymm0
vmovaps %ymm1, 0x620(%rsp)
vmovaps %ymm0, 0x600(%rsp)
vmovdqa 0x620(%rsp), %ymm0
vmovdqa 0x600(%rsp), %ymm1
vpand %ymm1, %ymm0, %ymm0
vmovdqa %ymm0, 0x20(%rsp)
vmovaps 0x120(%rsp), %ymm1
vmovaps 0xe0(%rsp), %ymm0
vmovaps %ymm1, 0xe20(%rsp)
vmovaps %ymm0, 0xe00(%rsp)
vmovdqa 0xe20(%rsp), %ymm0
vpternlogq $0xf, %ymm0, %ymm0, %ymm0
vmovaps 0xe00(%rsp), %ymm1
vpand %ymm1, %ymm0, %ymm0
vmovaps %ymm0, 0xe0(%rsp)
vmovaps 0xe0(%rsp), %ymm1
vmovaps 0x20(%rsp), %ymm0
vmovaps %ymm1, 0x740(%rsp)
vmovaps %ymm0, 0x720(%rsp)
vmovaps 0x740(%rsp), %ymm0
vaddps 0x720(%rsp), %ymm0, %ymm0
vmovaps %ymm0, 0xe0(%rsp)
vmovaps 0xe0(%rsp), %ymm1
vmovaps 0x100(%rsp), %ymm0
vmovaps %ymm1, 0xda0(%rsp)
vmovaps %ymm0, 0xd80(%rsp)
vmovaps 0xda0(%rsp), %ymm0
vmovaps 0xd80(%rsp), %ymm1
vpxor %ymm1, %ymm0, %ymm0
vmovaps %ymm0, 0xe0(%rsp)
vmovaps 0xe0(%rsp), %ymm0
movq %rbp, %rsp
popq %rbp
retq
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,291 |
ncnn::UnaryOp_x86_avx512_functor::unary_op_sin::func_pack4(float vector[4] const&) const
|
__m128 func_pack4(const __m128& x) const
{
return sin_ps(x);
}
|
subq $0x678, %rsp # imm = 0x678
movq %rdi, -0x78(%rsp)
movq %rsi, -0x80(%rsp)
movq -0x80(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm0, 0x40(%rsp)
vpxor %xmm0, %xmm0, %xmm0
vmovaps %xmm0, 0x70(%rsp)
vmovaps 0x70(%rsp), %xmm0
vmovaps %xmm0, 0x20(%rsp)
vmovaps 0x40(%rsp), %xmm0
vmovaps %xmm0, (%rsp)
vmovaps 0x40(%rsp), %xmm0
vmovaps %xmm0, 0x1d0(%rsp)
vbroadcastss 0x7ca2b7(%rip), %xmm0 # 0x1e1cdb0
vmovaps %xmm0, 0x1c0(%rsp)
vmovdqa 0x1d0(%rsp), %xmm0
vmovdqa 0x1c0(%rsp), %xmm1
vpand %xmm1, %xmm0, %xmm0
vmovdqa %xmm0, 0x40(%rsp)
vmovaps (%rsp), %xmm0
vmovaps %xmm0, 0x1b0(%rsp)
vbroadcastss 0x7bb383(%rip), %xmm0 # 0x1e0deb8
vmovaps %xmm0, 0x1a0(%rsp)
vmovdqa 0x1b0(%rsp), %xmm0
vmovdqa 0x1a0(%rsp), %xmm1
vpand %xmm1, %xmm0, %xmm0
vmovdqa %xmm0, (%rsp)
vmovaps 0x40(%rsp), %xmm0
vmovaps %xmm0, 0x110(%rsp)
vbroadcastss 0x7ca247(%rip), %xmm0 # 0x1e1cdb8
vmovaps %xmm0, 0x100(%rsp)
vmovaps 0x110(%rsp), %xmm0
vmovaps 0x100(%rsp), %xmm1
vmulps %xmm1, %xmm0, %xmm0
vmovaps %xmm0, -0x10(%rsp)
vmovaps -0x10(%rsp), %xmm0
vmovaps %xmm0, 0x160(%rsp)
vcvttps2dq 0x160(%rsp), %xmm0
vmovdqa %xmm0, -0x30(%rsp)
vmovdqa -0x30(%rsp), %xmm0
vmovdqa %xmm0, 0x3a0(%rsp)
vpbroadcastq 0x7ca214(%rip), %xmm0 # 0x1e1cde0
vmovdqa %xmm0, 0x390(%rsp)
vmovdqa 0x3a0(%rsp), %xmm0
vmovdqa 0x390(%rsp), %xmm1
vpaddd %xmm1, %xmm0, %xmm0
vmovdqa %xmm0, -0x30(%rsp)
vmovdqa -0x30(%rsp), %xmm0
vmovdqa %xmm0, 0x5f0(%rsp)
vpbroadcastq 0x7ca1df(%rip), %xmm0 # 0x1e1cde8
vmovdqa %xmm0, 0x5e0(%rsp)
vmovdqa 0x5f0(%rsp), %xmm0
vmovdqa 0x5e0(%rsp), %xmm1
vpand %xmm1, %xmm0, %xmm0
vmovdqa %xmm0, -0x30(%rsp)
vmovdqa -0x30(%rsp), %xmm0
vmovdqa %xmm0, 0x170(%rsp)
vcvtdq2ps 0x170(%rsp), %xmm0
vmovaps %xmm0, -0x10(%rsp)
vmovdqa -0x30(%rsp), %xmm0
vmovdqa %xmm0, 0x5d0(%rsp)
vpbroadcastq 0x7ca18c(%rip), %xmm0 # 0x1e1cdf0
vmovdqa %xmm0, 0x5c0(%rsp)
vmovdqa 0x5d0(%rsp), %xmm0
vmovdqa 0x5c0(%rsp), %xmm1
vpand %xmm1, %xmm0, %xmm0
vmovdqa %xmm0, -0x20(%rsp)
vmovdqa -0x20(%rsp), %xmm0
vmovdqa %xmm0, 0x3c0(%rsp)
movl $0x1d, 0x3bc(%rsp)
vmovdqa 0x3c0(%rsp), %xmm0
vmovd 0x3bc(%rsp), %xmm1
vpslld %xmm1, %xmm0, %xmm0
vmovdqa %xmm0, -0x20(%rsp)
vmovdqa -0x30(%rsp), %xmm0
vmovdqa %xmm0, 0x5b0(%rsp)
vpbroadcastq 0x7ca121(%rip), %xmm0 # 0x1e1cdf8
vmovdqa %xmm0, 0x5a0(%rsp)
vmovdqa 0x5b0(%rsp), %xmm0
vmovdqa 0x5a0(%rsp), %xmm1
vpand %xmm1, %xmm0, %xmm0
vmovdqa %xmm0, -0x30(%rsp)
vmovdqa -0x30(%rsp), %xmm1
vpxor %xmm0, %xmm0, %xmm0
vmovdqa %xmm0, 0x620(%rsp)
vmovdqa 0x620(%rsp), %xmm0
vmovdqa %xmm1, 0x610(%rsp)
vmovdqa %xmm0, 0x600(%rsp)
vmovdqa 0x610(%rsp), %xmm0
vmovdqa 0x600(%rsp), %xmm1
vpcmpeqd %xmm1, %xmm0, %xmm0
vmovdqa %xmm0, -0x30(%rsp)
vmovdqa -0x20(%rsp), %xmm0
vmovdqa %xmm0, 0x3e0(%rsp)
vmovdqa 0x3e0(%rsp), %xmm0
vmovdqa %xmm0, -0x40(%rsp)
vmovdqa -0x30(%rsp), %xmm0
vmovdqa %xmm0, 0x3d0(%rsp)
vmovdqa 0x3d0(%rsp), %xmm0
vmovdqa %xmm0, -0x50(%rsp)
vmovaps (%rsp), %xmm1
vmovaps -0x40(%rsp), %xmm0
vmovaps %xmm1, 0x660(%rsp)
vmovaps %xmm0, 0x650(%rsp)
vmovdqa 0x660(%rsp), %xmm0
vmovdqa 0x650(%rsp), %xmm1
vpxor %xmm1, %xmm0, %xmm0
vmovdqa %xmm0, (%rsp)
vbroadcastss 0x7c9ff9(%rip), %xmm0 # 0x1e1cdbc
vmovaps %xmm0, 0x30(%rsp)
vbroadcastss 0x7c9fee(%rip), %xmm0 # 0x1e1cdc0
vmovaps %xmm0, 0x20(%rsp)
vbroadcastss 0x7c9fe3(%rip), %xmm0 # 0x1e1cdc4
vmovaps %xmm0, 0x10(%rsp)
vmovaps -0x10(%rsp), %xmm2
vmovaps 0x30(%rsp), %xmm1
vmovaps 0x40(%rsp), %xmm0
vmovaps %xmm2, 0x380(%rsp)
vmovaps %xmm1, 0x370(%rsp)
vmovaps %xmm0, 0x360(%rsp)
vmovaps 0x380(%rsp), %xmm2
vmovaps 0x370(%rsp), %xmm1
vmovaps 0x360(%rsp), %xmm0
vmovaps %xmm2, 0x440(%rsp)
vmovaps %xmm1, 0x430(%rsp)
vmovaps %xmm0, 0x420(%rsp)
vmovaps 0x440(%rsp), %xmm1
vmovaps 0x430(%rsp), %xmm0
vmovaps 0x420(%rsp), %xmm2
vfmadd213ps %xmm2, %xmm1, %xmm0 # xmm0 = (xmm1 * xmm0) + xmm2
vmovaps %xmm0, 0x40(%rsp)
vmovaps -0x10(%rsp), %xmm2
vmovaps 0x20(%rsp), %xmm1
vmovaps 0x40(%rsp), %xmm0
vmovaps %xmm2, 0x350(%rsp)
vmovaps %xmm1, 0x340(%rsp)
vmovaps %xmm0, 0x330(%rsp)
vmovaps 0x350(%rsp), %xmm2
vmovaps 0x340(%rsp), %xmm1
vmovaps 0x330(%rsp), %xmm0
vmovaps %xmm2, 0x470(%rsp)
vmovaps %xmm1, 0x460(%rsp)
vmovaps %xmm0, 0x450(%rsp)
vmovaps 0x470(%rsp), %xmm1
vmovaps 0x460(%rsp), %xmm0
vmovaps 0x450(%rsp), %xmm2
vfmadd213ps %xmm2, %xmm1, %xmm0 # xmm0 = (xmm1 * xmm0) + xmm2
vmovaps %xmm0, 0x40(%rsp)
vmovaps -0x10(%rsp), %xmm2
vmovaps 0x10(%rsp), %xmm1
vmovaps 0x40(%rsp), %xmm0
vmovaps %xmm2, 0x320(%rsp)
vmovaps %xmm1, 0x310(%rsp)
vmovaps %xmm0, 0x300(%rsp)
vmovaps 0x320(%rsp), %xmm2
vmovaps 0x310(%rsp), %xmm1
vmovaps 0x300(%rsp), %xmm0
vmovaps %xmm2, 0x4a0(%rsp)
vmovaps %xmm1, 0x490(%rsp)
vmovaps %xmm0, 0x480(%rsp)
vmovaps 0x4a0(%rsp), %xmm1
vmovaps 0x490(%rsp), %xmm0
vmovaps 0x480(%rsp), %xmm2
vfmadd213ps %xmm2, %xmm1, %xmm0 # xmm0 = (xmm1 * xmm0) + xmm2
vmovaps %xmm0, 0x40(%rsp)
vbroadcastss 0x7c9e3d(%rip), %xmm0 # 0x1e1cdc8
vmovaps %xmm0, -0x10(%rsp)
vmovaps 0x40(%rsp), %xmm0
vmovaps %xmm0, 0xf0(%rsp)
vmovaps %xmm0, 0xe0(%rsp)
vmovaps 0xf0(%rsp), %xmm0
vmovaps 0xe0(%rsp), %xmm1
vmulps %xmm1, %xmm0, %xmm0
vmovaps %xmm0, -0x60(%rsp)
vmovaps -0x10(%rsp), %xmm1
vmovaps -0x60(%rsp), %xmm0
vmovaps %xmm1, 0x2f0(%rsp)
vmovaps %xmm0, 0x2e0(%rsp)
vbroadcastss 0x7c9de0(%rip), %xmm0 # 0x1e1cdcc
vmovaps %xmm0, 0x2d0(%rsp)
vmovaps 0x2f0(%rsp), %xmm2
vmovaps 0x2e0(%rsp), %xmm1
vmovaps 0x2d0(%rsp), %xmm0
vmovaps %xmm2, 0x4d0(%rsp)
vmovaps %xmm1, 0x4c0(%rsp)
vmovaps %xmm0, 0x4b0(%rsp)
vmovaps 0x4d0(%rsp), %xmm1
vmovaps 0x4c0(%rsp), %xmm0
vmovaps 0x4b0(%rsp), %xmm2
vfmadd213ps %xmm2, %xmm1, %xmm0 # xmm0 = (xmm1 * xmm0) + xmm2
vmovaps %xmm0, -0x10(%rsp)
vmovaps -0x10(%rsp), %xmm1
vmovaps -0x60(%rsp), %xmm0
vmovaps %xmm1, 0x2c0(%rsp)
vmovaps %xmm0, 0x2b0(%rsp)
vbroadcastss 0x7c9d58(%rip), %xmm0 # 0x1e1cdd0
vmovaps %xmm0, 0x2a0(%rsp)
vmovaps 0x2c0(%rsp), %xmm2
vmovaps 0x2b0(%rsp), %xmm1
vmovaps 0x2a0(%rsp), %xmm0
vmovaps %xmm2, 0x500(%rsp)
vmovaps %xmm1, 0x4f0(%rsp)
vmovaps %xmm0, 0x4e0(%rsp)
vmovaps 0x500(%rsp), %xmm1
vmovaps 0x4f0(%rsp), %xmm0
vmovaps 0x4e0(%rsp), %xmm2
vfmadd213ps %xmm2, %xmm1, %xmm0 # xmm0 = (xmm1 * xmm0) + xmm2
vmovaps %xmm0, -0x10(%rsp)
vmovaps -0x10(%rsp), %xmm1
vmovaps -0x60(%rsp), %xmm0
vmovaps %xmm1, 0xd0(%rsp)
vmovaps %xmm0, 0xc0(%rsp)
vmovaps 0xd0(%rsp), %xmm0
vmovaps 0xc0(%rsp), %xmm1
vmulps %xmm1, %xmm0, %xmm0
vmovaps %xmm0, -0x10(%rsp)
vmovaps -0x10(%rsp), %xmm1
vmovaps -0x60(%rsp), %xmm0
vmovaps %xmm1, 0xb0(%rsp)
vmovaps %xmm0, 0xa0(%rsp)
vmovaps 0xb0(%rsp), %xmm0
vmovaps 0xa0(%rsp), %xmm1
vmulps %xmm1, %xmm0, %xmm0
vmovaps %xmm0, -0x10(%rsp)
vmovaps -0x60(%rsp), %xmm1
vmovaps -0x10(%rsp), %xmm0
vmovaps %xmm1, 0x200(%rsp)
vbroadcastss 0x7b5ead(%rip), %xmm1 # 0x1e0901c
vmovaps %xmm1, 0x1f0(%rsp)
vmovaps %xmm0, 0x1e0(%rsp)
vmovaps 0x200(%rsp), %xmm2
vmovaps 0x1f0(%rsp), %xmm1
vmovaps 0x1e0(%rsp), %xmm0
vmovaps %xmm2, 0x410(%rsp)
vmovaps %xmm1, 0x400(%rsp)
vmovaps %xmm0, 0x3f0(%rsp)
vmovaps 0x410(%rsp), %xmm1
vmovaps 0x400(%rsp), %xmm0
vmovaps 0x3f0(%rsp), %xmm2
vfnmadd213ps %xmm2, %xmm1, %xmm0 # xmm0 = -(xmm1 * xmm0) + xmm2
vmovaps %xmm0, -0x10(%rsp)
vmovaps -0x10(%rsp), %xmm0
vmovaps %xmm0, 0x150(%rsp)
vbroadcastss 0x7b5e0f(%rip), %xmm0 # 0x1e09004
vmovaps %xmm0, 0x140(%rsp)
vmovaps 0x150(%rsp), %xmm0
vmovaps 0x140(%rsp), %xmm1
vaddps %xmm1, %xmm0, %xmm0
vmovaps %xmm0, -0x10(%rsp)
vbroadcastss 0x7c9bb1(%rip), %xmm0 # 0x1e1cdd4
vmovaps %xmm0, -0x70(%rsp)
vmovaps -0x70(%rsp), %xmm1
vmovaps -0x60(%rsp), %xmm0
vmovaps %xmm1, 0x290(%rsp)
vmovaps %xmm0, 0x280(%rsp)
vbroadcastss 0x7c9b88(%rip), %xmm0 # 0x1e1cdd8
vmovaps %xmm0, 0x270(%rsp)
vmovaps 0x290(%rsp), %xmm2
vmovaps 0x280(%rsp), %xmm1
vmovaps 0x270(%rsp), %xmm0
vmovaps %xmm2, 0x530(%rsp)
vmovaps %xmm1, 0x520(%rsp)
vmovaps %xmm0, 0x510(%rsp)
vmovaps 0x530(%rsp), %xmm1
vmovaps 0x520(%rsp), %xmm0
vmovaps 0x510(%rsp), %xmm2
vfmadd213ps %xmm2, %xmm1, %xmm0 # xmm0 = (xmm1 * xmm0) + xmm2
vmovaps %xmm0, -0x70(%rsp)
vmovaps -0x70(%rsp), %xmm1
vmovaps -0x60(%rsp), %xmm0
vmovaps %xmm1, 0x260(%rsp)
vmovaps %xmm0, 0x250(%rsp)
vbroadcastss 0x7c9b00(%rip), %xmm0 # 0x1e1cddc
vmovaps %xmm0, 0x240(%rsp)
vmovaps 0x260(%rsp), %xmm2
vmovaps 0x250(%rsp), %xmm1
vmovaps 0x240(%rsp), %xmm0
vmovaps %xmm2, 0x560(%rsp)
vmovaps %xmm1, 0x550(%rsp)
vmovaps %xmm0, 0x540(%rsp)
vmovaps 0x560(%rsp), %xmm1
vmovaps 0x550(%rsp), %xmm0
vmovaps 0x540(%rsp), %xmm2
vfmadd213ps %xmm2, %xmm1, %xmm0 # xmm0 = (xmm1 * xmm0) + xmm2
vmovaps %xmm0, -0x70(%rsp)
vmovaps -0x70(%rsp), %xmm1
vmovaps -0x60(%rsp), %xmm0
vmovaps %xmm1, 0x90(%rsp)
vmovaps %xmm0, 0x80(%rsp)
vmovaps 0x90(%rsp), %xmm0
vmovaps 0x80(%rsp), %xmm1
vmulps %xmm1, %xmm0, %xmm0
vmovaps %xmm0, -0x70(%rsp)
vmovaps -0x70(%rsp), %xmm1
vmovaps 0x40(%rsp), %xmm0
vmovaps %xmm1, 0x230(%rsp)
vmovaps %xmm0, 0x220(%rsp)
vmovaps %xmm0, 0x210(%rsp)
vmovaps 0x230(%rsp), %xmm2
vmovaps 0x220(%rsp), %xmm1
vmovaps 0x210(%rsp), %xmm0
vmovaps %xmm2, 0x590(%rsp)
vmovaps %xmm1, 0x580(%rsp)
vmovaps %xmm0, 0x570(%rsp)
vmovaps 0x590(%rsp), %xmm1
vmovaps 0x580(%rsp), %xmm0
vmovaps 0x570(%rsp), %xmm2
vfmadd213ps %xmm2, %xmm1, %xmm0 # xmm0 = (xmm1 * xmm0) + xmm2
vmovaps %xmm0, -0x70(%rsp)
vmovaps -0x50(%rsp), %xmm0
vmovaps %xmm0, 0x10(%rsp)
vmovaps 0x10(%rsp), %xmm1
vmovaps -0x70(%rsp), %xmm0
vmovaps %xmm1, 0x190(%rsp)
vmovaps %xmm0, 0x180(%rsp)
vmovdqa 0x190(%rsp), %xmm0
vmovdqa 0x180(%rsp), %xmm1
vpand %xmm1, %xmm0, %xmm0
vmovdqa %xmm0, -0x70(%rsp)
vmovaps 0x10(%rsp), %xmm1
vmovaps -0x10(%rsp), %xmm0
vmovaps %xmm1, 0x60(%rsp)
vmovaps %xmm0, 0x50(%rsp)
vmovdqa 0x60(%rsp), %xmm0
vpternlogq $0xf, %xmm0, %xmm0, %xmm0
vmovaps 0x50(%rsp), %xmm1
vpand %xmm1, %xmm0, %xmm0
vmovaps %xmm0, -0x10(%rsp)
vmovaps -0x10(%rsp), %xmm1
vmovaps -0x70(%rsp), %xmm0
vmovaps %xmm1, 0x130(%rsp)
vmovaps %xmm0, 0x120(%rsp)
vmovaps 0x130(%rsp), %xmm0
vaddps 0x120(%rsp), %xmm0, %xmm0
vmovaps %xmm0, -0x10(%rsp)
vmovaps -0x10(%rsp), %xmm1
vmovaps (%rsp), %xmm0
vmovaps %xmm1, 0x640(%rsp)
vmovaps %xmm0, 0x630(%rsp)
vmovaps 0x640(%rsp), %xmm0
vmovaps 0x630(%rsp), %xmm1
vpxor %xmm1, %xmm0, %xmm0
vmovaps %xmm0, -0x10(%rsp)
vmovaps -0x10(%rsp), %xmm0
addq $0x678, %rsp # imm = 0x678
retq
nopw %cs:(%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,292 |
ncnn::UnaryOp_x86_avx512_functor::unary_op_sin::func(float const&) const
|
float func(const float& x) const
{
return (float)sin(x);
}
|
subq $0x18, %rsp
movq %rdi, 0x10(%rsp)
movq %rsi, 0x8(%rsp)
movq 0x8(%rsp), %rax
vmovss (%rax), %xmm0
callq 0x163a260
addq $0x18, %rsp
retq
nopw %cs:(%rax,%rax)
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,293 |
ncnn::UnaryOp_x86_avx512_functor::unary_op_cos::func_pack16(float vector[16] const&) const
|
__m512 func_pack16(const __m512& x) const
{
return cos512_ps(x);
}
|
pushq %rbp
movq %rsp, %rbp
andq $-0x40, %rsp
subq $0x15c0, %rsp # imm = 0x15C0
movq %rdi, 0x38(%rsp)
movq %rsi, 0x30(%rsp)
movq 0x30(%rsp), %rax
vmovaps (%rax), %zmm0
vmovaps %zmm0, 0x2c0(%rsp)
vmovaps 0x2c0(%rsp), %zmm0
vmovaps %zmm0, 0x440(%rsp)
vmovaps 0x440(%rsp), %zmm1
movl $0x7fffffff, 0x4fc(%rsp) # imm = 0x7FFFFFFF
vpbroadcastd 0x4fc(%rsp), %zmm0
vmovdqa64 %zmm0, 0x480(%rsp)
vmovdqa64 0x480(%rsp), %zmm0
vmovaps %zmm1, 0x400(%rsp)
vmovdqa64 %zmm0, 0x3c0(%rsp)
vmovdqa64 0x400(%rsp), %zmm0
vmovdqa64 0x3c0(%rsp), %zmm1
vpandq %zmm1, %zmm0, %zmm0
vmovdqa64 %zmm0, 0x380(%rsp)
vmovdqa64 0x380(%rsp), %zmm0
vmovdqa64 %zmm0, 0x2c0(%rsp)
vmovaps 0x2c0(%rsp), %zmm1
vmovaps 0x7ca516(%rip), %zmm0 # 0x1e1db00
vmovaps %zmm1, 0x8c0(%rsp)
vmovaps %zmm0, 0x880(%rsp)
vmovaps 0x8c0(%rsp), %zmm0
vmovaps 0x880(%rsp), %zmm1
vmulps %zmm1, %zmm0, %zmm0
vmovaps %zmm0, 0x1c0(%rsp)
vmovaps 0x1c0(%rsp), %zmm0
vmovaps %zmm0, 0x1000(%rsp)
vpxor %xmm1, %xmm1, %xmm1
vcvttps2dq 0x1000(%rsp), %zmm0
vmovdqa64 %zmm1, 0x1140(%rsp)
vmovdqa64 %zmm0, 0x140(%rsp)
vmovdqa64 0x140(%rsp), %zmm1
vmovdqa64 0x7ca4ea(%rip), %zmm0 # 0x1e1db40
vmovdqa64 %zmm1, 0x1080(%rsp)
vmovdqa64 %zmm0, 0x1040(%rsp)
vmovdqa64 0x1080(%rsp), %zmm0
vmovdqa64 0x1040(%rsp), %zmm1
vpaddd %zmm1, %zmm0, %zmm0
vmovdqa64 %zmm0, 0x140(%rsp)
vmovdqa64 0x140(%rsp), %zmm1
vmovdqa64 0x7ca4ea(%rip), %zmm0 # 0x1e1db80
vmovdqa64 %zmm1, 0x13c0(%rsp)
vmovdqa64 %zmm0, 0x1380(%rsp)
vmovdqa64 0x13c0(%rsp), %zmm0
vmovdqa64 0x1380(%rsp), %zmm1
vpandq %zmm1, %zmm0, %zmm0
vmovdqa64 %zmm0, 0x140(%rsp)
vmovdqa64 0x140(%rsp), %zmm0
vmovdqa64 %zmm0, 0x1280(%rsp)
vcvtdq2ps 0x1280(%rsp), %zmm0
vmovaps %zmm0, 0x1c0(%rsp)
vmovdqa64 0x140(%rsp), %zmm1
vmovdqa64 0x7ca50a(%rip), %zmm0 # 0x1e1dc00
vmovdqa64 %zmm1, 0x1240(%rsp)
vmovdqa64 %zmm0, 0x1200(%rsp)
vmovdqa64 0x1240(%rsp), %zmm0
vmovdqa64 0x1200(%rsp), %zmm1
vpsubd %zmm1, %zmm0, %zmm0
vmovdqa64 %zmm0, 0x140(%rsp)
vmovdqa64 0x140(%rsp), %zmm1
vmovdqa64 0x7ca48a(%rip), %zmm0 # 0x1e1dbc0
vmovdqa64 %zmm1, 0x1540(%rsp)
vmovdqa64 %zmm0, 0x1500(%rsp)
vmovdqa64 0x1540(%rsp), %zmm0
vmovdqa64 0x1500(%rsp), %zmm1
vpandnq %zmm1, %zmm0, %zmm0
vmovdqa64 %zmm0, 0x180(%rsp)
vmovdqa64 0x180(%rsp), %zmm0
vmovdqa64 %zmm0, 0x1100(%rsp)
movl $0x1d, 0x10fc(%rsp)
vmovdqa64 0x1100(%rsp), %zmm0
vmovd 0x10fc(%rsp), %xmm1
vpslld %xmm1, %zmm0, %zmm0
vmovdqa64 %zmm0, 0x180(%rsp)
vmovdqa64 0x140(%rsp), %zmm1
vmovdqa64 0x7ca450(%rip), %zmm0 # 0x1e1dc00
vmovdqa64 %zmm1, 0x1340(%rsp)
vmovdqa64 %zmm0, 0x1300(%rsp)
vmovdqa64 0x1340(%rsp), %zmm0
vmovdqa64 0x1300(%rsp), %zmm1
vpandq %zmm1, %zmm0, %zmm0
vmovdqa64 %zmm0, 0x140(%rsp)
vmovdqa64 0x140(%rsp), %zmm0
vptestnmd %zmm0, %zmm0, %k0
kmovw %k0, 0x12fe(%rsp)
kmovw 0x12fe(%rsp), %k0
vpmovm2d %k0, %zmm0
vmovdqa64 %zmm0, 0x140(%rsp)
vmovdqa64 0x180(%rsp), %zmm0
vmovdqa64 %zmm0, 0x340(%rsp)
vmovdqa64 0x340(%rsp), %zmm0
vmovdqa64 %zmm0, 0x100(%rsp)
vmovdqa64 0x140(%rsp), %zmm0
vmovdqa64 %zmm0, 0x300(%rsp)
vmovdqa64 0x300(%rsp), %zmm0
vmovdqa64 %zmm0, 0xc0(%rsp)
vmovaps 0x7ca42a(%rip), %zmm0 # 0x1e1dc80
vmovaps %zmm0, 0x280(%rsp)
vmovaps 0x7ca458(%rip), %zmm0 # 0x1e1dcc0
vmovaps %zmm0, 0x240(%rsp)
vmovaps 0x7ca486(%rip), %zmm0 # 0x1e1dd00
vmovaps %zmm0, 0x200(%rsp)
vmovaps 0x1c0(%rsp), %zmm2
vmovaps 0x280(%rsp), %zmm1
vmovaps 0x2c0(%rsp), %zmm0
vmovaps %zmm2, 0xc80(%rsp)
vmovaps %zmm1, 0xc40(%rsp)
vmovaps %zmm0, 0xc00(%rsp)
vmovaps 0xc80(%rsp), %zmm1
vmovaps 0xc40(%rsp), %zmm0
vmovaps 0xc00(%rsp), %zmm2
vfmadd213ps %zmm2, %zmm1, %zmm0 # zmm0 = (zmm1 * zmm0) + zmm2
vmovaps %zmm0, 0x2c0(%rsp)
vmovaps 0x1c0(%rsp), %zmm2
vmovaps 0x240(%rsp), %zmm1
vmovaps 0x2c0(%rsp), %zmm0
vmovaps %zmm2, 0xbc0(%rsp)
vmovaps %zmm1, 0xb80(%rsp)
vmovaps %zmm0, 0xb40(%rsp)
vmovaps 0xbc0(%rsp), %zmm1
vmovaps 0xb80(%rsp), %zmm0
vmovaps 0xb40(%rsp), %zmm2
vfmadd213ps %zmm2, %zmm1, %zmm0 # zmm0 = (zmm1 * zmm0) + zmm2
vmovaps %zmm0, 0x2c0(%rsp)
vmovaps 0x1c0(%rsp), %zmm2
vmovaps 0x200(%rsp), %zmm1
vmovaps 0x2c0(%rsp), %zmm0
vmovaps %zmm2, 0xb00(%rsp)
vmovaps %zmm1, 0xac0(%rsp)
vmovaps %zmm0, 0xa80(%rsp)
vmovaps 0xb00(%rsp), %zmm1
vmovaps 0xac0(%rsp), %zmm0
vmovaps 0xa80(%rsp), %zmm2
vfmadd213ps %zmm2, %zmm1, %zmm0 # zmm0 = (zmm1 * zmm0) + zmm2
vmovaps %zmm0, 0x2c0(%rsp)
vmovaps 0x7ca3b2(%rip), %zmm0 # 0x1e1dd40
vmovaps %zmm0, 0x1c0(%rsp)
vmovaps 0x2c0(%rsp), %zmm0
vmovaps %zmm0, 0x840(%rsp)
vmovaps %zmm0, 0x800(%rsp)
vmovaps 0x840(%rsp), %zmm0
vmovaps 0x800(%rsp), %zmm1
vmulps %zmm1, %zmm0, %zmm0
vmovaps %zmm0, 0x80(%rsp)
vmovaps 0x1c0(%rsp), %zmm2
vmovaps 0x80(%rsp), %zmm1
vmovaps 0x7ca39a(%rip), %zmm0 # 0x1e1dd80
vmovaps %zmm2, 0xa40(%rsp)
vmovaps %zmm1, 0xa00(%rsp)
vmovaps %zmm0, 0x9c0(%rsp)
vmovaps 0xa40(%rsp), %zmm1
vmovaps 0xa00(%rsp), %zmm0
vmovaps 0x9c0(%rsp), %zmm2
vfmadd213ps %zmm2, %zmm1, %zmm0 # zmm0 = (zmm1 * zmm0) + zmm2
vmovaps %zmm0, 0x1c0(%rsp)
vmovaps 0x1c0(%rsp), %zmm2
vmovaps 0x80(%rsp), %zmm1
vmovaps 0x7ca382(%rip), %zmm0 # 0x1e1ddc0
vmovaps %zmm2, 0x980(%rsp)
vmovaps %zmm1, 0x940(%rsp)
vmovaps %zmm0, 0x900(%rsp)
vmovaps 0x980(%rsp), %zmm1
vmovaps 0x940(%rsp), %zmm0
vmovaps 0x900(%rsp), %zmm2
vfmadd213ps %zmm2, %zmm1, %zmm0 # zmm0 = (zmm1 * zmm0) + zmm2
vmovaps %zmm0, 0x1c0(%rsp)
vmovaps 0x1c0(%rsp), %zmm1
vmovaps 0x80(%rsp), %zmm0
vmovaps %zmm1, 0x7c0(%rsp)
vmovaps %zmm0, 0x780(%rsp)
vmovaps 0x7c0(%rsp), %zmm0
vmovaps 0x780(%rsp), %zmm1
vmulps %zmm1, %zmm0, %zmm0
vmovaps %zmm0, 0x1c0(%rsp)
vmovaps 0x1c0(%rsp), %zmm1
vmovaps 0x80(%rsp), %zmm0
vmovaps %zmm1, 0x740(%rsp)
vmovaps %zmm0, 0x700(%rsp)
vmovaps 0x740(%rsp), %zmm0
vmovaps 0x700(%rsp), %zmm1
vmulps %zmm1, %zmm0, %zmm0
vmovaps %zmm0, 0x1c0(%rsp)
vmovaps 0x80(%rsp), %zmm2
vmovaps 0x7c9436(%rip), %zmm1 # 0x1e1cf40
vmovaps 0x1c0(%rsp), %zmm0
vmovaps %zmm2, 0xd40(%rsp)
vmovaps %zmm1, 0xd00(%rsp)
vmovaps %zmm0, 0xcc0(%rsp)
vmovaps 0xd40(%rsp), %zmm1
vmovaps 0xd00(%rsp), %zmm0
vmovaps 0xcc0(%rsp), %zmm2
vfnmadd213ps %zmm2, %zmm1, %zmm0 # zmm0 = -(zmm1 * zmm0) + zmm2
vmovaps %zmm0, 0x1c0(%rsp)
vmovaps 0x1c0(%rsp), %zmm1
vmovaps 0x7c92de(%rip), %zmm0 # 0x1e1ce40
vmovaps %zmm1, 0xfc0(%rsp)
vmovaps %zmm0, 0xf80(%rsp)
vmovaps 0xfc0(%rsp), %zmm0
vmovaps 0xf80(%rsp), %zmm1
vaddps %zmm1, %zmm0, %zmm0
vmovaps %zmm0, 0x1c0(%rsp)
vmovaps 0x7ca266(%rip), %zmm0 # 0x1e1de00
vmovaps %zmm0, 0x40(%rsp)
vmovaps 0x40(%rsp), %zmm1
vmovaps 0x80(%rsp), %zmm0
vmovaps %zmm1, 0x6c0(%rsp)
vmovaps %zmm0, 0x680(%rsp)
vmovaps 0x6c0(%rsp), %zmm0
vmovaps 0x680(%rsp), %zmm1
vmulps %zmm1, %zmm0, %zmm0
vmovaps %zmm0, 0x40(%rsp)
vmovaps 0x40(%rsp), %zmm1
vmovaps 0x7ca24e(%rip), %zmm0 # 0x1e1de40
vmovaps %zmm1, 0xf40(%rsp)
vmovaps %zmm0, 0xf00(%rsp)
vmovaps 0xf40(%rsp), %zmm0
vmovaps 0xf00(%rsp), %zmm1
vaddps %zmm1, %zmm0, %zmm0
vmovaps %zmm0, 0x40(%rsp)
vmovaps 0x40(%rsp), %zmm1
vmovaps 0x80(%rsp), %zmm0
vmovaps %zmm1, 0x640(%rsp)
vmovaps %zmm0, 0x600(%rsp)
vmovaps 0x640(%rsp), %zmm0
vmovaps 0x600(%rsp), %zmm1
vmulps %zmm1, %zmm0, %zmm0
vmovaps %zmm0, 0x40(%rsp)
vmovaps 0x40(%rsp), %zmm1
vmovaps 0x7ca210(%rip), %zmm0 # 0x1e1de80
vmovaps %zmm1, 0xec0(%rsp)
vmovaps %zmm0, 0xe80(%rsp)
vmovaps 0xec0(%rsp), %zmm0
vmovaps 0xe80(%rsp), %zmm1
vaddps %zmm1, %zmm0, %zmm0
vmovaps %zmm0, 0x40(%rsp)
vmovaps 0x40(%rsp), %zmm1
vmovaps 0x80(%rsp), %zmm0
vmovaps %zmm1, 0x5c0(%rsp)
vmovaps %zmm0, 0x580(%rsp)
vmovaps 0x5c0(%rsp), %zmm0
vmovaps 0x580(%rsp), %zmm1
vmulps %zmm1, %zmm0, %zmm0
vmovaps %zmm0, 0x40(%rsp)
vmovaps 0x40(%rsp), %zmm1
vmovaps 0x2c0(%rsp), %zmm0
vmovaps %zmm1, 0x540(%rsp)
vmovaps %zmm0, 0x500(%rsp)
vmovaps 0x540(%rsp), %zmm0
vmovaps 0x500(%rsp), %zmm1
vmulps %zmm1, %zmm0, %zmm0
vmovaps %zmm0, 0x40(%rsp)
vmovaps 0x40(%rsp), %zmm1
vmovaps 0x2c0(%rsp), %zmm0
vmovaps %zmm1, 0xe40(%rsp)
vmovaps %zmm0, 0xe00(%rsp)
vmovaps 0xe40(%rsp), %zmm0
vmovaps 0xe00(%rsp), %zmm1
vaddps %zmm1, %zmm0, %zmm0
vmovaps %zmm0, 0x40(%rsp)
vmovaps 0xc0(%rsp), %zmm0
vmovaps %zmm0, 0x200(%rsp)
vmovaps 0x200(%rsp), %zmm1
vmovaps 0x40(%rsp), %zmm0
vmovaps %zmm1, 0x11c0(%rsp)
vmovaps %zmm0, 0x1180(%rsp)
vmovdqa64 0x11c0(%rsp), %zmm0
vmovdqa64 0x1180(%rsp), %zmm1
vpandd %zmm1, %zmm0, %zmm0
vmovdqa64 %zmm0, 0x40(%rsp)
vmovaps 0x200(%rsp), %zmm1
vmovaps 0x1c0(%rsp), %zmm0
vmovaps %zmm1, 0x14c0(%rsp)
vmovaps %zmm0, 0x1480(%rsp)
vmovdqa64 0x14c0(%rsp), %zmm0
vpternlogq $0xf, %zmm0, %zmm0, %zmm0
vmovaps 0x1480(%rsp), %zmm1
vpandd %zmm1, %zmm0, %zmm0
vmovaps %zmm0, 0x1c0(%rsp)
vmovaps 0x1c0(%rsp), %zmm1
vmovaps 0x40(%rsp), %zmm0
vmovaps %zmm1, 0xdc0(%rsp)
vmovaps %zmm0, 0xd80(%rsp)
vmovaps 0xdc0(%rsp), %zmm0
vaddps 0xd80(%rsp), %zmm0, %zmm0
vmovaps %zmm0, 0x1c0(%rsp)
vmovaps 0x1c0(%rsp), %zmm1
vmovaps 0x100(%rsp), %zmm0
vmovaps %zmm1, 0x1440(%rsp)
vmovaps %zmm0, 0x1400(%rsp)
vmovaps 0x1440(%rsp), %zmm0
vmovaps 0x1400(%rsp), %zmm1
vpxord %zmm1, %zmm0, %zmm0
vmovaps %zmm0, 0x1c0(%rsp)
vmovaps 0x1c0(%rsp), %zmm0
movq %rbp, %rsp
popq %rbp
retq
nop
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,294 |
ncnn::UnaryOp_x86_avx512_functor::unary_op_cos::func_pack8(float vector[8] const&) const
|
__m256 func_pack8(const __m256& x) const
{
return cos256_ps(x);
}
|
pushq %rbp
movq %rsp, %rbp
andq $-0x20, %rsp
subq $0xd80, %rsp # imm = 0xD80
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
movq 0x10(%rsp), %rax
vmovaps (%rax), %ymm0
vmovaps %ymm0, 0x160(%rsp)
vpxor %xmm0, %xmm0, %xmm0
vmovaps %ymm0, 0x180(%rsp)
vmovaps 0x180(%rsp), %ymm0
vmovaps %ymm0, 0x120(%rsp)
vmovaps 0x160(%rsp), %ymm0
vmovaps %ymm0, 0x5e0(%rsp)
vbroadcastss 0x7c8edb(%rip), %ymm0 # 0x1e1cdb0
vmovaps %ymm0, 0x5c0(%rsp)
vmovdqa 0x5e0(%rsp), %ymm0
vmovdqa 0x5c0(%rsp), %ymm1
vpand %ymm1, %ymm0, %ymm0
vmovdqa %ymm0, 0x160(%rsp)
vmovaps 0x160(%rsp), %ymm0
vmovaps %ymm0, 0x380(%rsp)
vbroadcastss 0x7c8ea0(%rip), %ymm0 # 0x1e1cdb8
vmovaps %ymm0, 0x360(%rsp)
vmovaps 0x380(%rsp), %ymm0
vmovaps 0x360(%rsp), %ymm1
vmulps %ymm1, %ymm0, %ymm0
vmovaps %ymm0, 0xe0(%rsp)
vmovaps 0xe0(%rsp), %ymm0
vmovaps %ymm0, 0x7a0(%rsp)
vcvttps2dq 0x7a0(%rsp), %ymm0
vmovdqa %ymm0, 0xa0(%rsp)
vmovdqa 0xa0(%rsp), %ymm0
vmovdqa %ymm0, 0x7e0(%rsp)
vpbroadcastq 0x7c8e61(%rip), %ymm0 # 0x1e1cde0
vmovdqa %ymm0, 0x7c0(%rsp)
vmovdqa 0x7e0(%rsp), %ymm1
vmovdqa 0x7c0(%rsp), %ymm0
vmovdqa %ymm1, 0xae0(%rsp)
vmovdqa %ymm0, 0xac0(%rsp)
vmovdqa 0xae0(%rsp), %ymm0
vmovdqa 0xac0(%rsp), %ymm1
vpaddd %ymm1, %ymm0, %ymm0
vmovdqa %ymm0, 0xa0(%rsp)
vmovdqa 0xa0(%rsp), %ymm0
vmovdqa %ymm0, 0xc40(%rsp)
vpbroadcastq 0x7c8e02(%rip), %ymm0 # 0x1e1cde8
vmovdqa %ymm0, 0xc20(%rsp)
vmovdqa 0xc40(%rsp), %ymm0
vmovdqa 0xc20(%rsp), %ymm1
vpand %ymm1, %ymm0, %ymm0
vmovdqa %ymm0, 0xa0(%rsp)
vmovdqa 0xa0(%rsp), %ymm0
vmovdqa %ymm0, 0xb80(%rsp)
vcvtdq2ps 0xb80(%rsp), %ymm0
vmovaps %ymm0, 0xe0(%rsp)
vmovdqa 0xa0(%rsp), %ymm0
vmovdqa %ymm0, 0xb60(%rsp)
vpbroadcastq 0x7c8dab(%rip), %ymm0 # 0x1e1cdf8
vmovdqa %ymm0, 0xb40(%rsp)
vmovdqa 0xb60(%rsp), %ymm2
vmovdqa 0xb40(%rsp), %ymm1
vmovdqa %ymm2, 0xbc0(%rsp)
vmovdqa %ymm1, 0xba0(%rsp)
vmovdqa 0xbc0(%rsp), %ymm1
vmovdqa 0xba0(%rsp), %ymm2
vpsubd %ymm2, %ymm1, %ymm1
vmovdqa %ymm1, 0xa0(%rsp)
vmovdqa 0xa0(%rsp), %ymm1
vmovdqa %ymm1, 0xd40(%rsp)
vpbroadcastq 0x7c8d3c(%rip), %ymm1 # 0x1e1cdf0
vmovdqa %ymm1, 0xd20(%rsp)
vmovdqa 0xd40(%rsp), %ymm1
vmovdqa 0xd20(%rsp), %ymm2
vpandn %ymm2, %ymm1, %ymm1
vmovdqa %ymm1, 0xc0(%rsp)
vmovdqa 0xc0(%rsp), %ymm1
vmovdqa %ymm1, 0x820(%rsp)
movl $0x1d, 0x81c(%rsp)
vmovdqa 0x820(%rsp), %ymm1
movl 0x81c(%rsp), %eax
vmovdqa %ymm1, 0xb20(%rsp)
movl %eax, 0xb1c(%rsp)
vmovdqa 0xb20(%rsp), %ymm1
vmovd 0xb1c(%rsp), %xmm2
vpslld %xmm2, %ymm1, %ymm1
vmovdqa %ymm1, 0xc0(%rsp)
vmovdqa 0xa0(%rsp), %ymm1
vmovdqa %ymm1, 0xc00(%rsp)
vmovdqa %ymm0, 0xbe0(%rsp)
vmovdqa 0xc00(%rsp), %ymm0
vmovdqa 0xbe0(%rsp), %ymm1
vpand %ymm1, %ymm0, %ymm0
vmovdqa %ymm0, 0xa0(%rsp)
vmovdqa 0xa0(%rsp), %ymm0
vmovdqa %ymm0, 0xc80(%rsp)
vpxor %xmm0, %xmm0, %xmm0
vmovdqa %ymm0, 0xc60(%rsp)
vmovdqa 0xc80(%rsp), %ymm0
vmovdqa 0xc60(%rsp), %ymm1
vpcmpeqd %ymm1, %ymm0, %ymm0
vmovdqa %ymm0, 0xa0(%rsp)
vmovdqa 0xc0(%rsp), %ymm0
vmovdqa %ymm0, 0x860(%rsp)
vmovdqa 0x860(%rsp), %ymm0
vmovdqa %ymm0, 0x80(%rsp)
vmovdqa 0xa0(%rsp), %ymm0
vmovdqa %ymm0, 0x840(%rsp)
vmovdqa 0x840(%rsp), %ymm0
vmovdqa %ymm0, 0x60(%rsp)
vbroadcastss 0x7c8bbe(%rip), %ymm0 # 0x1e1cdbc
vmovaps %ymm0, 0x140(%rsp)
vbroadcastss 0x7c8bb0(%rip), %ymm0 # 0x1e1cdc0
vmovaps %ymm0, 0x120(%rsp)
vbroadcastss 0x7c8ba2(%rip), %ymm0 # 0x1e1cdc4
vmovaps %ymm0, 0x100(%rsp)
vmovaps 0xe0(%rsp), %ymm2
vmovaps 0x140(%rsp), %ymm1
vmovaps 0x160(%rsp), %ymm0
vmovaps %ymm2, 0x560(%rsp)
vmovaps %ymm1, 0x540(%rsp)
vmovaps %ymm0, 0x520(%rsp)
vmovaps 0x560(%rsp), %ymm2
vmovaps 0x540(%rsp), %ymm1
vmovaps 0x520(%rsp), %ymm0
vmovaps %ymm2, 0x8c0(%rsp)
vmovaps %ymm1, 0x8a0(%rsp)
vmovaps %ymm0, 0x880(%rsp)
vmovaps 0x8c0(%rsp), %ymm1
vmovaps 0x8a0(%rsp), %ymm0
vmovaps 0x880(%rsp), %ymm2
vfmadd213ps %ymm2, %ymm1, %ymm0 # ymm0 = (ymm1 * ymm0) + ymm2
vmovaps %ymm0, 0x160(%rsp)
vmovaps 0xe0(%rsp), %ymm2
vmovaps 0x120(%rsp), %ymm1
vmovaps 0x160(%rsp), %ymm0
vmovaps %ymm2, 0x500(%rsp)
vmovaps %ymm1, 0x4e0(%rsp)
vmovaps %ymm0, 0x4c0(%rsp)
vmovaps 0x500(%rsp), %ymm2
vmovaps 0x4e0(%rsp), %ymm1
vmovaps 0x4c0(%rsp), %ymm0
vmovaps %ymm2, 0x920(%rsp)
vmovaps %ymm1, 0x900(%rsp)
vmovaps %ymm0, 0x8e0(%rsp)
vmovaps 0x920(%rsp), %ymm1
vmovaps 0x900(%rsp), %ymm0
vmovaps 0x8e0(%rsp), %ymm2
vfmadd213ps %ymm2, %ymm1, %ymm0 # ymm0 = (ymm1 * ymm0) + ymm2
vmovaps %ymm0, 0x160(%rsp)
vmovaps 0xe0(%rsp), %ymm2
vmovaps 0x100(%rsp), %ymm1
vmovaps 0x160(%rsp), %ymm0
vmovaps %ymm2, 0x4a0(%rsp)
vmovaps %ymm1, 0x480(%rsp)
vmovaps %ymm0, 0x460(%rsp)
vmovaps 0x4a0(%rsp), %ymm2
vmovaps 0x480(%rsp), %ymm1
vmovaps 0x460(%rsp), %ymm0
vmovaps %ymm2, 0x980(%rsp)
vmovaps %ymm1, 0x960(%rsp)
vmovaps %ymm0, 0x940(%rsp)
vmovaps 0x980(%rsp), %ymm1
vmovaps 0x960(%rsp), %ymm0
vmovaps 0x940(%rsp), %ymm2
vfmadd213ps %ymm2, %ymm1, %ymm0 # ymm0 = (ymm1 * ymm0) + ymm2
vmovaps %ymm0, 0x160(%rsp)
vbroadcastss 0x7c89d5(%rip), %ymm0 # 0x1e1cdc8
vmovaps %ymm0, 0xe0(%rsp)
vmovaps 0x160(%rsp), %ymm0
vmovaps %ymm0, 0x340(%rsp)
vmovaps %ymm0, 0x320(%rsp)
vmovaps 0x340(%rsp), %ymm0
vmovaps 0x320(%rsp), %ymm1
vmulps %ymm1, %ymm0, %ymm0
vmovaps %ymm0, 0x40(%rsp)
vmovaps 0xe0(%rsp), %ymm1
vmovaps 0x40(%rsp), %ymm0
vmovaps %ymm1, 0x440(%rsp)
vmovaps %ymm0, 0x420(%rsp)
vbroadcastss 0x7c896f(%rip), %ymm0 # 0x1e1cdcc
vmovaps %ymm0, 0x400(%rsp)
vmovaps 0x440(%rsp), %ymm2
vmovaps 0x420(%rsp), %ymm1
vmovaps 0x400(%rsp), %ymm0
vmovaps %ymm2, 0x9e0(%rsp)
vmovaps %ymm1, 0x9c0(%rsp)
vmovaps %ymm0, 0x9a0(%rsp)
vmovaps 0x9e0(%rsp), %ymm1
vmovaps 0x9c0(%rsp), %ymm0
vmovaps 0x9a0(%rsp), %ymm2
vfmadd213ps %ymm2, %ymm1, %ymm0 # ymm0 = (ymm1 * ymm0) + ymm2
vmovaps %ymm0, 0xe0(%rsp)
vmovaps 0xe0(%rsp), %ymm1
vmovaps 0x40(%rsp), %ymm0
vmovaps %ymm1, 0x3e0(%rsp)
vmovaps %ymm0, 0x3c0(%rsp)
vbroadcastss 0x7c88e1(%rip), %ymm0 # 0x1e1cdd0
vmovaps %ymm0, 0x3a0(%rsp)
vmovaps 0x3e0(%rsp), %ymm2
vmovaps 0x3c0(%rsp), %ymm1
vmovaps 0x3a0(%rsp), %ymm0
vmovaps %ymm2, 0xa40(%rsp)
vmovaps %ymm1, 0xa20(%rsp)
vmovaps %ymm0, 0xa00(%rsp)
vmovaps 0xa40(%rsp), %ymm1
vmovaps 0xa20(%rsp), %ymm0
vmovaps 0xa00(%rsp), %ymm2
vfmadd213ps %ymm2, %ymm1, %ymm0 # ymm0 = (ymm1 * ymm0) + ymm2
vmovaps %ymm0, 0xe0(%rsp)
vmovaps 0xe0(%rsp), %ymm1
vmovaps 0x40(%rsp), %ymm0
vmovaps %ymm1, 0x300(%rsp)
vmovaps %ymm0, 0x2e0(%rsp)
vmovaps 0x300(%rsp), %ymm0
vmovaps 0x2e0(%rsp), %ymm1
vmulps %ymm1, %ymm0, %ymm0
vmovaps %ymm0, 0xe0(%rsp)
vmovaps 0xe0(%rsp), %ymm1
vmovaps 0x40(%rsp), %ymm0
vmovaps %ymm1, 0x2c0(%rsp)
vmovaps %ymm0, 0x2a0(%rsp)
vmovaps 0x2c0(%rsp), %ymm0
vmovaps 0x2a0(%rsp), %ymm1
vmulps %ymm1, %ymm0, %ymm0
vmovaps %ymm0, 0xe0(%rsp)
vmovaps 0x40(%rsp), %ymm1
vmovaps 0xe0(%rsp), %ymm0
vmovaps %ymm1, 0x640(%rsp)
vbroadcastss 0x7b4a24(%rip), %ymm1 # 0x1e0901c
vmovaps %ymm1, 0x620(%rsp)
vmovaps %ymm0, 0x600(%rsp)
vmovaps 0x640(%rsp), %ymm2
vmovaps 0x620(%rsp), %ymm1
vmovaps 0x600(%rsp), %ymm0
vmovaps %ymm2, 0xaa0(%rsp)
vmovaps %ymm1, 0xa80(%rsp)
vmovaps %ymm0, 0xa60(%rsp)
vmovaps 0xaa0(%rsp), %ymm1
vmovaps 0xa80(%rsp), %ymm0
vmovaps 0xa60(%rsp), %ymm2
vfnmadd213ps %ymm2, %ymm1, %ymm0 # ymm0 = -(ymm1 * ymm0) + ymm2
vmovaps %ymm0, 0xe0(%rsp)
vmovaps 0xe0(%rsp), %ymm0
vmovaps %ymm0, 0x780(%rsp)
vbroadcastss 0x7b4980(%rip), %ymm0 # 0x1e09004
vmovaps %ymm0, 0x760(%rsp)
vmovaps 0x780(%rsp), %ymm0
vmovaps 0x760(%rsp), %ymm1
vaddps %ymm1, %ymm0, %ymm0
vmovaps %ymm0, 0xe0(%rsp)
vbroadcastss 0x7c871f(%rip), %ymm0 # 0x1e1cdd4
vmovaps %ymm0, 0x20(%rsp)
vmovaps 0x20(%rsp), %ymm1
vmovaps 0x40(%rsp), %ymm0
vmovaps %ymm1, 0x280(%rsp)
vmovaps %ymm0, 0x260(%rsp)
vmovaps 0x280(%rsp), %ymm0
vmovaps 0x260(%rsp), %ymm1
vmulps %ymm1, %ymm0, %ymm0
vmovaps %ymm0, 0x20(%rsp)
vmovaps 0x20(%rsp), %ymm0
vmovaps %ymm0, 0x740(%rsp)
vbroadcastss 0x7c86cb(%rip), %ymm0 # 0x1e1cdd8
vmovaps %ymm0, 0x720(%rsp)
vmovaps 0x740(%rsp), %ymm0
vmovaps 0x720(%rsp), %ymm1
vaddps %ymm1, %ymm0, %ymm0
vmovaps %ymm0, 0x20(%rsp)
vmovaps 0x20(%rsp), %ymm1
vmovaps 0x40(%rsp), %ymm0
vmovaps %ymm1, 0x240(%rsp)
vmovaps %ymm0, 0x220(%rsp)
vmovaps 0x240(%rsp), %ymm0
vmovaps 0x220(%rsp), %ymm1
vmulps %ymm1, %ymm0, %ymm0
vmovaps %ymm0, 0x20(%rsp)
vmovaps 0x20(%rsp), %ymm0
vmovaps %ymm0, 0x700(%rsp)
vbroadcastss 0x7c8658(%rip), %ymm0 # 0x1e1cddc
vmovaps %ymm0, 0x6e0(%rsp)
vmovaps 0x700(%rsp), %ymm0
vmovaps 0x6e0(%rsp), %ymm1
vaddps %ymm1, %ymm0, %ymm0
vmovaps %ymm0, 0x20(%rsp)
vmovaps 0x20(%rsp), %ymm1
vmovaps 0x40(%rsp), %ymm0
vmovaps %ymm1, 0x200(%rsp)
vmovaps %ymm0, 0x1e0(%rsp)
vmovaps 0x200(%rsp), %ymm0
vmovaps 0x1e0(%rsp), %ymm1
vmulps %ymm1, %ymm0, %ymm0
vmovaps %ymm0, 0x20(%rsp)
vmovaps 0x20(%rsp), %ymm1
vmovaps 0x160(%rsp), %ymm0
vmovaps %ymm1, 0x1c0(%rsp)
vmovaps %ymm0, 0x1a0(%rsp)
vmovaps 0x1c0(%rsp), %ymm0
vmovaps 0x1a0(%rsp), %ymm1
vmulps %ymm1, %ymm0, %ymm0
vmovaps %ymm0, 0x20(%rsp)
vmovaps 0x20(%rsp), %ymm1
vmovaps 0x160(%rsp), %ymm0
vmovaps %ymm1, 0x6c0(%rsp)
vmovaps %ymm0, 0x6a0(%rsp)
vmovaps 0x6c0(%rsp), %ymm0
vmovaps 0x6a0(%rsp), %ymm1
vaddps %ymm1, %ymm0, %ymm0
vmovaps %ymm0, 0x20(%rsp)
vmovaps 0x60(%rsp), %ymm0
vmovaps %ymm0, 0x100(%rsp)
vmovaps 0x100(%rsp), %ymm1
vmovaps 0x20(%rsp), %ymm0
vmovaps %ymm1, 0x5a0(%rsp)
vmovaps %ymm0, 0x580(%rsp)
vmovdqa 0x5a0(%rsp), %ymm0
vmovdqa 0x580(%rsp), %ymm1
vpand %ymm1, %ymm0, %ymm0
vmovdqa %ymm0, 0x20(%rsp)
vmovaps 0x100(%rsp), %ymm1
vmovaps 0xe0(%rsp), %ymm0
vmovaps %ymm1, 0xd00(%rsp)
vmovaps %ymm0, 0xce0(%rsp)
vmovdqa 0xd00(%rsp), %ymm0
vpternlogq $0xf, %ymm0, %ymm0, %ymm0
vmovaps 0xce0(%rsp), %ymm1
vpand %ymm1, %ymm0, %ymm0
vmovaps %ymm0, 0xe0(%rsp)
vmovaps 0xe0(%rsp), %ymm1
vmovaps 0x20(%rsp), %ymm0
vmovaps %ymm1, 0x680(%rsp)
vmovaps %ymm0, 0x660(%rsp)
vmovaps 0x680(%rsp), %ymm0
vaddps 0x660(%rsp), %ymm0, %ymm0
vmovaps %ymm0, 0xe0(%rsp)
vmovaps 0xe0(%rsp), %ymm1
vmovaps 0x80(%rsp), %ymm0
vmovaps %ymm1, 0xcc0(%rsp)
vmovaps %ymm0, 0xca0(%rsp)
vmovaps 0xcc0(%rsp), %ymm0
vmovaps 0xca0(%rsp), %ymm1
vpxor %ymm1, %ymm0, %ymm0
vmovaps %ymm0, 0xe0(%rsp)
vmovaps 0xe0(%rsp), %ymm0
movq %rbp, %rsp
popq %rbp
retq
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,295 |
ncnn::UnaryOp_x86_avx512_functor::unary_op_cos::func_pack4(float vector[4] const&) const
|
__m128 func_pack4(const __m128& x) const
{
return cos_ps(x);
}
|
subq $0x648, %rsp # imm = 0x648
movq %rdi, -0x78(%rsp)
movq %rsi, -0x80(%rsp)
movq -0x80(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm0, 0x30(%rsp)
vpxor %xmm0, %xmm0, %xmm0
vmovaps %xmm0, 0x60(%rsp)
vmovaps 0x60(%rsp), %xmm0
vmovaps %xmm0, 0x10(%rsp)
vmovaps 0x30(%rsp), %xmm0
vmovaps %xmm0, 0x1a0(%rsp)
vbroadcastss 0x7c83e2(%rip), %xmm0 # 0x1e1cdb0
vmovaps %xmm0, 0x190(%rsp)
vmovdqa 0x1a0(%rsp), %xmm0
vmovdqa 0x190(%rsp), %xmm1
vpand %xmm1, %xmm0, %xmm0
vmovdqa %xmm0, 0x30(%rsp)
vmovaps 0x30(%rsp), %xmm0
vmovaps %xmm0, 0x100(%rsp)
vbroadcastss 0x7c83ad(%rip), %xmm0 # 0x1e1cdb8
vmovaps %xmm0, 0xf0(%rsp)
vmovaps 0x100(%rsp), %xmm0
vmovaps 0xf0(%rsp), %xmm1
vmulps %xmm1, %xmm0, %xmm0
vmovaps %xmm0, -0x10(%rsp)
vmovaps -0x10(%rsp), %xmm0
vmovaps %xmm0, 0x150(%rsp)
vcvttps2dq 0x150(%rsp), %xmm0
vmovdqa %xmm0, -0x30(%rsp)
vmovdqa -0x30(%rsp), %xmm0
vmovdqa %xmm0, 0x370(%rsp)
vpbroadcastq 0x7c837a(%rip), %xmm0 # 0x1e1cde0
vmovdqa %xmm0, 0x360(%rsp)
vmovdqa 0x370(%rsp), %xmm0
vmovdqa 0x360(%rsp), %xmm1
vpaddd %xmm1, %xmm0, %xmm0
vmovdqa %xmm0, -0x30(%rsp)
vmovdqa -0x30(%rsp), %xmm0
vmovdqa %xmm0, 0x5c0(%rsp)
vpbroadcastq 0x7c8345(%rip), %xmm0 # 0x1e1cde8
vmovdqa %xmm0, 0x5b0(%rsp)
vmovdqa 0x5c0(%rsp), %xmm0
vmovdqa 0x5b0(%rsp), %xmm1
vpand %xmm1, %xmm0, %xmm0
vmovdqa %xmm0, -0x30(%rsp)
vmovdqa -0x30(%rsp), %xmm0
vmovdqa %xmm0, 0x160(%rsp)
vcvtdq2ps 0x160(%rsp), %xmm0
vmovaps %xmm0, -0x10(%rsp)
vmovdqa -0x30(%rsp), %xmm0
vmovdqa %xmm0, 0x580(%rsp)
vpbroadcastq 0x7c82fa(%rip), %xmm0 # 0x1e1cdf8
vmovdqa %xmm0, 0x570(%rsp)
vmovdqa 0x580(%rsp), %xmm1
vmovdqa 0x570(%rsp), %xmm2
vpsubd %xmm2, %xmm1, %xmm1
vmovdqa %xmm1, -0x30(%rsp)
vmovdqa -0x30(%rsp), %xmm1
vmovdqa %xmm1, 0x630(%rsp)
vpbroadcastq 0x7c82b5(%rip), %xmm1 # 0x1e1cdf0
vmovdqa %xmm1, 0x620(%rsp)
vmovdqa 0x630(%rsp), %xmm1
vmovdqa 0x620(%rsp), %xmm2
vpandn %xmm2, %xmm1, %xmm1
vmovdqa %xmm1, -0x20(%rsp)
vmovdqa -0x20(%rsp), %xmm1
vmovdqa %xmm1, 0x390(%rsp)
movl $0x1d, 0x38c(%rsp)
vmovdqa 0x390(%rsp), %xmm1
vmovd 0x38c(%rsp), %xmm2
vpslld %xmm2, %xmm1, %xmm1
vmovdqa %xmm1, -0x20(%rsp)
vmovdqa -0x30(%rsp), %xmm1
vmovdqa %xmm1, 0x5a0(%rsp)
vmovdqa %xmm0, 0x590(%rsp)
vmovdqa 0x5a0(%rsp), %xmm0
vmovdqa 0x590(%rsp), %xmm1
vpand %xmm1, %xmm0, %xmm0
vmovdqa %xmm0, -0x30(%rsp)
vmovdqa -0x30(%rsp), %xmm1
vpxor %xmm0, %xmm0, %xmm0
vmovdqa %xmm0, 0x5f0(%rsp)
vmovdqa 0x5f0(%rsp), %xmm0
vmovdqa %xmm1, 0x5e0(%rsp)
vmovdqa %xmm0, 0x5d0(%rsp)
vmovdqa 0x5e0(%rsp), %xmm0
vmovdqa 0x5d0(%rsp), %xmm1
vpcmpeqd %xmm1, %xmm0, %xmm0
vmovdqa %xmm0, -0x30(%rsp)
vmovdqa -0x20(%rsp), %xmm0
vmovdqa %xmm0, 0x3b0(%rsp)
vmovdqa 0x3b0(%rsp), %xmm0
vmovdqa %xmm0, -0x40(%rsp)
vmovdqa -0x30(%rsp), %xmm0
vmovdqa %xmm0, 0x3a0(%rsp)
vmovdqa 0x3a0(%rsp), %xmm0
vmovdqa %xmm0, -0x50(%rsp)
vbroadcastss 0x7c8163(%rip), %xmm0 # 0x1e1cdbc
vmovaps %xmm0, 0x20(%rsp)
vbroadcastss 0x7c8158(%rip), %xmm0 # 0x1e1cdc0
vmovaps %xmm0, 0x10(%rsp)
vbroadcastss 0x7c814d(%rip), %xmm0 # 0x1e1cdc4
vmovaps %xmm0, (%rsp)
vmovaps -0x10(%rsp), %xmm2
vmovaps 0x20(%rsp), %xmm1
vmovaps 0x30(%rsp), %xmm0
vmovaps %xmm2, 0x350(%rsp)
vmovaps %xmm1, 0x340(%rsp)
vmovaps %xmm0, 0x330(%rsp)
vmovaps 0x350(%rsp), %xmm2
vmovaps 0x340(%rsp), %xmm1
vmovaps 0x330(%rsp), %xmm0
vmovaps %xmm2, 0x410(%rsp)
vmovaps %xmm1, 0x400(%rsp)
vmovaps %xmm0, 0x3f0(%rsp)
vmovaps 0x410(%rsp), %xmm1
vmovaps 0x400(%rsp), %xmm0
vmovaps 0x3f0(%rsp), %xmm2
vfmadd213ps %xmm2, %xmm1, %xmm0 # xmm0 = (xmm1 * xmm0) + xmm2
vmovaps %xmm0, 0x30(%rsp)
vmovaps -0x10(%rsp), %xmm2
vmovaps 0x10(%rsp), %xmm1
vmovaps 0x30(%rsp), %xmm0
vmovaps %xmm2, 0x320(%rsp)
vmovaps %xmm1, 0x310(%rsp)
vmovaps %xmm0, 0x300(%rsp)
vmovaps 0x320(%rsp), %xmm2
vmovaps 0x310(%rsp), %xmm1
vmovaps 0x300(%rsp), %xmm0
vmovaps %xmm2, 0x440(%rsp)
vmovaps %xmm1, 0x430(%rsp)
vmovaps %xmm0, 0x420(%rsp)
vmovaps 0x440(%rsp), %xmm1
vmovaps 0x430(%rsp), %xmm0
vmovaps 0x420(%rsp), %xmm2
vfmadd213ps %xmm2, %xmm1, %xmm0 # xmm0 = (xmm1 * xmm0) + xmm2
vmovaps %xmm0, 0x30(%rsp)
vmovaps -0x10(%rsp), %xmm2
vmovaps (%rsp), %xmm1
vmovaps 0x30(%rsp), %xmm0
vmovaps %xmm2, 0x2f0(%rsp)
vmovaps %xmm1, 0x2e0(%rsp)
vmovaps %xmm0, 0x2d0(%rsp)
vmovaps 0x2f0(%rsp), %xmm2
vmovaps 0x2e0(%rsp), %xmm1
vmovaps 0x2d0(%rsp), %xmm0
vmovaps %xmm2, 0x470(%rsp)
vmovaps %xmm1, 0x460(%rsp)
vmovaps %xmm0, 0x450(%rsp)
vmovaps 0x470(%rsp), %xmm1
vmovaps 0x460(%rsp), %xmm0
vmovaps 0x450(%rsp), %xmm2
vfmadd213ps %xmm2, %xmm1, %xmm0 # xmm0 = (xmm1 * xmm0) + xmm2
vmovaps %xmm0, 0x30(%rsp)
vbroadcastss 0x7c7fa9(%rip), %xmm0 # 0x1e1cdc8
vmovaps %xmm0, -0x10(%rsp)
vmovaps 0x30(%rsp), %xmm0
vmovaps %xmm0, 0xe0(%rsp)
vmovaps %xmm0, 0xd0(%rsp)
vmovaps 0xe0(%rsp), %xmm0
vmovaps 0xd0(%rsp), %xmm1
vmulps %xmm1, %xmm0, %xmm0
vmovaps %xmm0, -0x60(%rsp)
vmovaps -0x10(%rsp), %xmm1
vmovaps -0x60(%rsp), %xmm0
vmovaps %xmm1, 0x2c0(%rsp)
vmovaps %xmm0, 0x2b0(%rsp)
vbroadcastss 0x7c7f4c(%rip), %xmm0 # 0x1e1cdcc
vmovaps %xmm0, 0x2a0(%rsp)
vmovaps 0x2c0(%rsp), %xmm2
vmovaps 0x2b0(%rsp), %xmm1
vmovaps 0x2a0(%rsp), %xmm0
vmovaps %xmm2, 0x4a0(%rsp)
vmovaps %xmm1, 0x490(%rsp)
vmovaps %xmm0, 0x480(%rsp)
vmovaps 0x4a0(%rsp), %xmm1
vmovaps 0x490(%rsp), %xmm0
vmovaps 0x480(%rsp), %xmm2
vfmadd213ps %xmm2, %xmm1, %xmm0 # xmm0 = (xmm1 * xmm0) + xmm2
vmovaps %xmm0, -0x10(%rsp)
vmovaps -0x10(%rsp), %xmm1
vmovaps -0x60(%rsp), %xmm0
vmovaps %xmm1, 0x290(%rsp)
vmovaps %xmm0, 0x280(%rsp)
vbroadcastss 0x7c7ec4(%rip), %xmm0 # 0x1e1cdd0
vmovaps %xmm0, 0x270(%rsp)
vmovaps 0x290(%rsp), %xmm2
vmovaps 0x280(%rsp), %xmm1
vmovaps 0x270(%rsp), %xmm0
vmovaps %xmm2, 0x4d0(%rsp)
vmovaps %xmm1, 0x4c0(%rsp)
vmovaps %xmm0, 0x4b0(%rsp)
vmovaps 0x4d0(%rsp), %xmm1
vmovaps 0x4c0(%rsp), %xmm0
vmovaps 0x4b0(%rsp), %xmm2
vfmadd213ps %xmm2, %xmm1, %xmm0 # xmm0 = (xmm1 * xmm0) + xmm2
vmovaps %xmm0, -0x10(%rsp)
vmovaps -0x10(%rsp), %xmm1
vmovaps -0x60(%rsp), %xmm0
vmovaps %xmm1, 0xc0(%rsp)
vmovaps %xmm0, 0xb0(%rsp)
vmovaps 0xc0(%rsp), %xmm0
vmovaps 0xb0(%rsp), %xmm1
vmulps %xmm1, %xmm0, %xmm0
vmovaps %xmm0, -0x10(%rsp)
vmovaps -0x10(%rsp), %xmm1
vmovaps -0x60(%rsp), %xmm0
vmovaps %xmm1, 0xa0(%rsp)
vmovaps %xmm0, 0x90(%rsp)
vmovaps 0xa0(%rsp), %xmm0
vmovaps 0x90(%rsp), %xmm1
vmulps %xmm1, %xmm0, %xmm0
vmovaps %xmm0, -0x10(%rsp)
vmovaps -0x60(%rsp), %xmm1
vmovaps -0x10(%rsp), %xmm0
vmovaps %xmm1, 0x1d0(%rsp)
vbroadcastss 0x7b4019(%rip), %xmm1 # 0x1e0901c
vmovaps %xmm1, 0x1c0(%rsp)
vmovaps %xmm0, 0x1b0(%rsp)
vmovaps 0x1d0(%rsp), %xmm2
vmovaps 0x1c0(%rsp), %xmm1
vmovaps 0x1b0(%rsp), %xmm0
vmovaps %xmm2, 0x3e0(%rsp)
vmovaps %xmm1, 0x3d0(%rsp)
vmovaps %xmm0, 0x3c0(%rsp)
vmovaps 0x3e0(%rsp), %xmm1
vmovaps 0x3d0(%rsp), %xmm0
vmovaps 0x3c0(%rsp), %xmm2
vfnmadd213ps %xmm2, %xmm1, %xmm0 # xmm0 = -(xmm1 * xmm0) + xmm2
vmovaps %xmm0, -0x10(%rsp)
vmovaps -0x10(%rsp), %xmm0
vmovaps %xmm0, 0x140(%rsp)
vbroadcastss 0x7b3f7b(%rip), %xmm0 # 0x1e09004
vmovaps %xmm0, 0x130(%rsp)
vmovaps 0x140(%rsp), %xmm0
vmovaps 0x130(%rsp), %xmm1
vaddps %xmm1, %xmm0, %xmm0
vmovaps %xmm0, -0x10(%rsp)
vbroadcastss 0x7c7d1d(%rip), %xmm0 # 0x1e1cdd4
vmovaps %xmm0, -0x70(%rsp)
vmovaps -0x70(%rsp), %xmm1
vmovaps -0x60(%rsp), %xmm0
vmovaps %xmm1, 0x260(%rsp)
vmovaps %xmm0, 0x250(%rsp)
vbroadcastss 0x7c7cf4(%rip), %xmm0 # 0x1e1cdd8
vmovaps %xmm0, 0x240(%rsp)
vmovaps 0x260(%rsp), %xmm2
vmovaps 0x250(%rsp), %xmm1
vmovaps 0x240(%rsp), %xmm0
vmovaps %xmm2, 0x500(%rsp)
vmovaps %xmm1, 0x4f0(%rsp)
vmovaps %xmm0, 0x4e0(%rsp)
vmovaps 0x500(%rsp), %xmm1
vmovaps 0x4f0(%rsp), %xmm0
vmovaps 0x4e0(%rsp), %xmm2
vfmadd213ps %xmm2, %xmm1, %xmm0 # xmm0 = (xmm1 * xmm0) + xmm2
vmovaps %xmm0, -0x70(%rsp)
vmovaps -0x70(%rsp), %xmm1
vmovaps -0x60(%rsp), %xmm0
vmovaps %xmm1, 0x230(%rsp)
vmovaps %xmm0, 0x220(%rsp)
vbroadcastss 0x7c7c6c(%rip), %xmm0 # 0x1e1cddc
vmovaps %xmm0, 0x210(%rsp)
vmovaps 0x230(%rsp), %xmm2
vmovaps 0x220(%rsp), %xmm1
vmovaps 0x210(%rsp), %xmm0
vmovaps %xmm2, 0x530(%rsp)
vmovaps %xmm1, 0x520(%rsp)
vmovaps %xmm0, 0x510(%rsp)
vmovaps 0x530(%rsp), %xmm1
vmovaps 0x520(%rsp), %xmm0
vmovaps 0x510(%rsp), %xmm2
vfmadd213ps %xmm2, %xmm1, %xmm0 # xmm0 = (xmm1 * xmm0) + xmm2
vmovaps %xmm0, -0x70(%rsp)
vmovaps -0x70(%rsp), %xmm1
vmovaps -0x60(%rsp), %xmm0
vmovaps %xmm1, 0x80(%rsp)
vmovaps %xmm0, 0x70(%rsp)
vmovaps 0x80(%rsp), %xmm0
vmovaps 0x70(%rsp), %xmm1
vmulps %xmm1, %xmm0, %xmm0
vmovaps %xmm0, -0x70(%rsp)
vmovaps -0x70(%rsp), %xmm1
vmovaps 0x30(%rsp), %xmm0
vmovaps %xmm1, 0x200(%rsp)
vmovaps %xmm0, 0x1f0(%rsp)
vmovaps %xmm0, 0x1e0(%rsp)
vmovaps 0x200(%rsp), %xmm2
vmovaps 0x1f0(%rsp), %xmm1
vmovaps 0x1e0(%rsp), %xmm0
vmovaps %xmm2, 0x560(%rsp)
vmovaps %xmm1, 0x550(%rsp)
vmovaps %xmm0, 0x540(%rsp)
vmovaps 0x560(%rsp), %xmm1
vmovaps 0x550(%rsp), %xmm0
vmovaps 0x540(%rsp), %xmm2
vfmadd213ps %xmm2, %xmm1, %xmm0 # xmm0 = (xmm1 * xmm0) + xmm2
vmovaps %xmm0, -0x70(%rsp)
vmovaps -0x50(%rsp), %xmm0
vmovaps %xmm0, (%rsp)
vmovaps (%rsp), %xmm1
vmovaps -0x70(%rsp), %xmm0
vmovaps %xmm1, 0x180(%rsp)
vmovaps %xmm0, 0x170(%rsp)
vmovdqa 0x180(%rsp), %xmm0
vmovdqa 0x170(%rsp), %xmm1
vpand %xmm1, %xmm0, %xmm0
vmovdqa %xmm0, -0x70(%rsp)
vmovaps (%rsp), %xmm1
vmovaps -0x10(%rsp), %xmm0
vmovaps %xmm1, 0x50(%rsp)
vmovaps %xmm0, 0x40(%rsp)
vmovdqa 0x50(%rsp), %xmm0
vpternlogq $0xf, %xmm0, %xmm0, %xmm0
vmovaps 0x40(%rsp), %xmm1
vpand %xmm1, %xmm0, %xmm0
vmovaps %xmm0, -0x10(%rsp)
vmovaps -0x10(%rsp), %xmm1
vmovaps -0x70(%rsp), %xmm0
vmovaps %xmm1, 0x120(%rsp)
vmovaps %xmm0, 0x110(%rsp)
vmovaps 0x120(%rsp), %xmm0
vaddps 0x110(%rsp), %xmm0, %xmm0
vmovaps %xmm0, -0x10(%rsp)
vmovaps -0x10(%rsp), %xmm1
vmovaps -0x40(%rsp), %xmm0
vmovaps %xmm1, 0x610(%rsp)
vmovaps %xmm0, 0x600(%rsp)
vmovaps 0x610(%rsp), %xmm0
vmovaps 0x600(%rsp), %xmm1
vpxor %xmm1, %xmm0, %xmm0
vmovaps %xmm0, -0x10(%rsp)
vmovaps -0x10(%rsp), %xmm0
addq $0x648, %rsp # imm = 0x648
retq
nopw %cs:(%rax,%rax)
nopl (%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,296 |
ncnn::UnaryOp_x86_avx512_functor::unary_op_cos::func(float const&) const
|
float func(const float& x) const
{
return (float)cos(x);
}
|
subq $0x18, %rsp
movq %rdi, 0x10(%rsp)
movq %rsi, 0x8(%rsp)
movq 0x8(%rsp), %rax
vmovss (%rax), %xmm0
callq 0x163a2b0
addq $0x18, %rsp
retq
nopw %cs:(%rax,%rax)
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,297 |
ncnn::UnaryOp_x86_avx512_functor::unary_op_tan::func_pack16(float vector[16] const&) const
|
__m512 func_pack16(const __m512& x) const
{
return tan512_ps(x);
}
|
pushq %rbp
movq %rsp, %rbp
andq $-0x40, %rsp
subq $0x1f00, %rsp # imm = 0x1F00
movq %rdi, 0x38(%rsp)
movq %rsi, 0x30(%rsp)
movq 0x30(%rsp), %rax
vmovaps (%rax), %zmm0
vmovaps %zmm0, 0x180(%rsp)
movl $0x322bcc77, 0x3fc(%rsp) # imm = 0x322BCC77
vbroadcastss 0x3fc(%rsp), %zmm0
vmovaps %zmm0, 0x380(%rsp)
vmovaps 0x380(%rsp), %zmm0
vmovaps %zmm0, 0xc0(%rsp)
vmovaps 0x180(%rsp), %zmm0
vmovaps %zmm0, 0x800(%rsp)
leaq 0x140(%rsp), %rax
movq %rax, 0x7f8(%rsp)
leaq 0x100(%rsp), %rax
movq %rax, 0x7f0(%rsp)
vmovaps 0x800(%rsp), %zmm0
vmovaps %zmm0, 0x6c0(%rsp)
vmovaps 0x800(%rsp), %zmm0
vmovaps %zmm0, 0x9c0(%rsp)
vmovaps 0x9c0(%rsp), %zmm1
movl $0x7fffffff, 0xa7c(%rsp) # imm = 0x7FFFFFFF
vpbroadcastd 0xa7c(%rsp), %zmm0
vmovdqa64 %zmm0, 0xa00(%rsp)
vmovdqa64 0xa00(%rsp), %zmm0
vmovaps %zmm1, 0x980(%rsp)
vmovdqa64 %zmm0, 0x940(%rsp)
vmovdqa64 0x980(%rsp), %zmm0
vmovdqa64 0x940(%rsp), %zmm1
vpandq %zmm1, %zmm0, %zmm0
vmovdqa64 %zmm0, 0x900(%rsp)
vmovdqa64 0x900(%rsp), %zmm0
vmovdqa64 %zmm0, 0x800(%rsp)
vmovaps 0x6c0(%rsp), %zmm1
vmovaps 0x7c85d8(%rip), %zmm0 # 0x1e1dac0
vmovaps %zmm1, 0x1900(%rsp)
vmovaps %zmm0, 0x18c0(%rsp)
vmovdqa64 0x1900(%rsp), %zmm0
vmovdqa64 0x18c0(%rsp), %zmm1
vpandd %zmm1, %zmm0, %zmm0
vmovdqa64 %zmm0, 0x6c0(%rsp)
vmovaps 0x800(%rsp), %zmm1
vmovaps 0x7c85d8(%rip), %zmm0 # 0x1e1db00
vmovaps %zmm1, 0xdc0(%rsp)
vmovaps %zmm0, 0xd80(%rsp)
vmovaps 0xdc0(%rsp), %zmm0
vmovaps 0xd80(%rsp), %zmm1
vmulps %zmm1, %zmm0, %zmm0
vmovaps %zmm0, 0x680(%rsp)
vmovaps 0x680(%rsp), %zmm0
vmovaps %zmm0, 0x1640(%rsp)
vpxor %xmm1, %xmm1, %xmm1
vcvttps2dq 0x1640(%rsp), %zmm0
vmovdqa64 %zmm1, 0x1800(%rsp)
vmovdqa64 %zmm0, 0x600(%rsp)
vmovdqa64 0x600(%rsp), %zmm1
vmovdqa64 0x7c85ac(%rip), %zmm0 # 0x1e1db40
vmovdqa64 %zmm1, 0x16c0(%rsp)
vmovdqa64 %zmm0, 0x1680(%rsp)
vmovdqa64 0x16c0(%rsp), %zmm0
vmovdqa64 0x1680(%rsp), %zmm1
vpaddd %zmm1, %zmm0, %zmm0
vmovdqa64 %zmm0, 0x600(%rsp)
vmovdqa64 0x600(%rsp), %zmm1
vmovdqa64 0x7c85ac(%rip), %zmm0 # 0x1e1db80
vmovdqa64 %zmm1, 0x1b80(%rsp)
vmovdqa64 %zmm0, 0x1b40(%rsp)
vmovdqa64 0x1b80(%rsp), %zmm0
vmovdqa64 0x1b40(%rsp), %zmm1
vpandq %zmm1, %zmm0, %zmm0
vmovdqa64 %zmm0, 0x600(%rsp)
vmovdqa64 0x600(%rsp), %zmm0
vmovdqa64 %zmm0, 0x19c0(%rsp)
vcvtdq2ps 0x19c0(%rsp), %zmm0
vmovaps %zmm0, 0x680(%rsp)
vmovdqa64 0x600(%rsp), %zmm0
vmovdqa64 %zmm0, 0x5c0(%rsp)
vmovdqa64 0x600(%rsp), %zmm1
vmovdqa64 0x7c857c(%rip), %zmm0 # 0x1e1dbc0
vmovdqa64 %zmm1, 0x1b00(%rsp)
vmovdqa64 %zmm0, 0x1ac0(%rsp)
vmovdqa64 0x1b00(%rsp), %zmm0
vmovdqa64 0x1ac0(%rsp), %zmm1
vpandq %zmm1, %zmm0, %zmm0
vmovdqa64 %zmm0, 0x640(%rsp)
vmovdqa64 0x640(%rsp), %zmm0
vmovdqa64 %zmm0, 0x17c0(%rsp)
movl $0x1d, 0x17bc(%rsp)
vmovdqa64 0x17c0(%rsp), %zmm0
vmovd 0x17bc(%rsp), %xmm1
vpslld %xmm1, %zmm0, %zmm0
vmovdqa64 %zmm0, 0x640(%rsp)
vmovdqa64 0x600(%rsp), %zmm1
vmovdqa64 0x7c8542(%rip), %zmm0 # 0x1e1dc00
vmovdqa64 %zmm1, 0x1a80(%rsp)
vmovdqa64 %zmm0, 0x1a40(%rsp)
vmovdqa64 0x1a80(%rsp), %zmm0
vmovdqa64 0x1a40(%rsp), %zmm1
vpandq %zmm1, %zmm0, %zmm0
vmovdqa64 %zmm0, 0x600(%rsp)
vmovdqa64 0x600(%rsp), %zmm0
vptestnmd %zmm0, %zmm0, %k0
kmovw %k0, 0x1a3e(%rsp)
kmovw 0x1a3e(%rsp), %k0
vpmovm2d %k0, %zmm0
vmovdqa64 %zmm0, 0x600(%rsp)
vmovdqa64 0x640(%rsp), %zmm0
vmovdqa64 %zmm0, 0x8c0(%rsp)
vmovdqa64 0x8c0(%rsp), %zmm0
vmovdqa64 %zmm0, 0x580(%rsp)
vmovdqa64 0x600(%rsp), %zmm0
vmovdqa64 %zmm0, 0x880(%rsp)
vmovdqa64 0x880(%rsp), %zmm0
vmovdqa64 %zmm0, 0x540(%rsp)
vmovaps 0x7c851c(%rip), %zmm0 # 0x1e1dc80
vmovaps %zmm0, 0x780(%rsp)
vmovaps 0x7c854a(%rip), %zmm0 # 0x1e1dcc0
vmovaps %zmm0, 0x740(%rsp)
vmovaps 0x7c8578(%rip), %zmm0 # 0x1e1dd00
vmovaps %zmm0, 0x700(%rsp)
vmovaps 0x680(%rsp), %zmm2
vmovaps 0x780(%rsp), %zmm1
vmovaps 0x800(%rsp), %zmm0
vmovaps %zmm2, 0x13c0(%rsp)
vmovaps %zmm1, 0x1380(%rsp)
vmovaps %zmm0, 0x1340(%rsp)
vmovaps 0x13c0(%rsp), %zmm1
vmovaps 0x1380(%rsp), %zmm0
vmovaps 0x1340(%rsp), %zmm2
vfmadd213ps %zmm2, %zmm1, %zmm0 # zmm0 = (zmm1 * zmm0) + zmm2
vmovaps %zmm0, 0x800(%rsp)
vmovaps 0x680(%rsp), %zmm2
vmovaps 0x740(%rsp), %zmm1
vmovaps 0x800(%rsp), %zmm0
vmovaps %zmm2, 0x1300(%rsp)
vmovaps %zmm1, 0x12c0(%rsp)
vmovaps %zmm0, 0x1280(%rsp)
vmovaps 0x1300(%rsp), %zmm1
vmovaps 0x12c0(%rsp), %zmm0
vmovaps 0x1280(%rsp), %zmm2
vfmadd213ps %zmm2, %zmm1, %zmm0 # zmm0 = (zmm1 * zmm0) + zmm2
vmovaps %zmm0, 0x800(%rsp)
vmovaps 0x680(%rsp), %zmm2
vmovaps 0x700(%rsp), %zmm1
vmovaps 0x800(%rsp), %zmm0
vmovaps %zmm2, 0x1240(%rsp)
vmovaps %zmm1, 0x1200(%rsp)
vmovaps %zmm0, 0x11c0(%rsp)
vmovaps 0x1240(%rsp), %zmm1
vmovaps 0x1200(%rsp), %zmm0
vmovaps 0x11c0(%rsp), %zmm2
vfmadd213ps %zmm2, %zmm1, %zmm0 # zmm0 = (zmm1 * zmm0) + zmm2
vmovaps %zmm0, 0x800(%rsp)
vmovdqa64 0x5c0(%rsp), %zmm1
vmovdqa64 0x7c835c(%rip), %zmm0 # 0x1e1dc00
vmovdqa64 %zmm1, 0x1980(%rsp)
vmovdqa64 %zmm0, 0x1940(%rsp)
vmovdqa64 0x1980(%rsp), %zmm0
vmovdqa64 0x1940(%rsp), %zmm1
vpsubd %zmm1, %zmm0, %zmm0
vmovdqa64 %zmm0, 0x5c0(%rsp)
vmovdqa64 0x5c0(%rsp), %zmm1
vmovdqa64 0x7c82dc(%rip), %zmm0 # 0x1e1dbc0
vmovdqa64 %zmm1, 0x1e00(%rsp)
vmovdqa64 %zmm0, 0x1dc0(%rsp)
vmovdqa64 0x1e00(%rsp), %zmm0
vmovdqa64 0x1dc0(%rsp), %zmm1
vpandnq %zmm1, %zmm0, %zmm0
vmovdqa64 %zmm0, 0x5c0(%rsp)
vmovdqa64 0x5c0(%rsp), %zmm0
vmovdqa64 %zmm0, 0x1740(%rsp)
movl $0x1d, 0x173c(%rsp)
vmovdqa64 0x1740(%rsp), %zmm0
vmovd 0x173c(%rsp), %xmm1
vpslld %xmm1, %zmm0, %zmm0
vmovdqa64 %zmm0, 0x5c0(%rsp)
vmovdqa64 0x5c0(%rsp), %zmm0
vmovdqa64 %zmm0, 0x840(%rsp)
vmovdqa64 0x840(%rsp), %zmm0
vmovdqa64 %zmm0, 0x500(%rsp)
vmovaps 0x6c0(%rsp), %zmm1
vmovaps 0x580(%rsp), %zmm0
vmovaps %zmm1, 0x1d00(%rsp)
vmovaps %zmm0, 0x1cc0(%rsp)
vmovdqa64 0x1d00(%rsp), %zmm0
vmovdqa64 0x1cc0(%rsp), %zmm1
vpxord %zmm1, %zmm0, %zmm0
vmovdqa64 %zmm0, 0x6c0(%rsp)
vmovaps 0x800(%rsp), %zmm0
vmovaps %zmm0, 0xd40(%rsp)
vmovaps %zmm0, 0xd00(%rsp)
vmovaps 0xd40(%rsp), %zmm0
vmovaps 0xd00(%rsp), %zmm1
vmulps %zmm1, %zmm0, %zmm0
vmovaps %zmm0, 0x4c0(%rsp)
vmovaps 0x7c8356(%rip), %zmm0 # 0x1e1dd40
vmovaps %zmm0, 0x680(%rsp)
vmovaps 0x680(%rsp), %zmm2
vmovaps 0x4c0(%rsp), %zmm1
vmovaps 0x7c8374(%rip), %zmm0 # 0x1e1dd80
vmovaps %zmm2, 0x1180(%rsp)
vmovaps %zmm1, 0x1140(%rsp)
vmovaps %zmm0, 0x1100(%rsp)
vmovaps 0x1180(%rsp), %zmm1
vmovaps 0x1140(%rsp), %zmm0
vmovaps 0x1100(%rsp), %zmm2
vfmadd213ps %zmm2, %zmm1, %zmm0 # zmm0 = (zmm1 * zmm0) + zmm2
vmovaps %zmm0, 0x680(%rsp)
vmovaps 0x680(%rsp), %zmm2
vmovaps 0x4c0(%rsp), %zmm1
vmovaps 0x7c835c(%rip), %zmm0 # 0x1e1ddc0
vmovaps %zmm2, 0x10c0(%rsp)
vmovaps %zmm1, 0x1080(%rsp)
vmovaps %zmm0, 0x1040(%rsp)
vmovaps 0x10c0(%rsp), %zmm1
vmovaps 0x1080(%rsp), %zmm0
vmovaps 0x1040(%rsp), %zmm2
vfmadd213ps %zmm2, %zmm1, %zmm0 # zmm0 = (zmm1 * zmm0) + zmm2
vmovaps %zmm0, 0x680(%rsp)
vmovaps 0x680(%rsp), %zmm1
vmovaps 0x4c0(%rsp), %zmm0
vmovaps %zmm1, 0xcc0(%rsp)
vmovaps %zmm0, 0xc80(%rsp)
vmovaps 0xcc0(%rsp), %zmm0
vmovaps 0xc80(%rsp), %zmm1
vmulps %zmm1, %zmm0, %zmm0
vmovaps %zmm0, 0x680(%rsp)
vmovaps 0x680(%rsp), %zmm1
vmovaps 0x4c0(%rsp), %zmm0
vmovaps %zmm1, 0xc40(%rsp)
vmovaps %zmm0, 0xc00(%rsp)
vmovaps 0xc40(%rsp), %zmm0
vmovaps 0xc00(%rsp), %zmm1
vmulps %zmm1, %zmm0, %zmm0
vmovaps %zmm0, 0x680(%rsp)
vmovaps 0x4c0(%rsp), %zmm2
vmovaps 0x7c7410(%rip), %zmm1 # 0x1e1cf40
vmovaps 0x680(%rsp), %zmm0
vmovaps %zmm2, 0x1480(%rsp)
vmovaps %zmm1, 0x1440(%rsp)
vmovaps %zmm0, 0x1400(%rsp)
vmovaps 0x1480(%rsp), %zmm1
vmovaps 0x1440(%rsp), %zmm0
vmovaps 0x1400(%rsp), %zmm2
vfnmadd213ps %zmm2, %zmm1, %zmm0 # zmm0 = -(zmm1 * zmm0) + zmm2
vmovaps %zmm0, 0x680(%rsp)
vmovaps 0x680(%rsp), %zmm1
vmovaps 0x7c72b8(%rip), %zmm0 # 0x1e1ce40
vmovaps %zmm1, 0x1600(%rsp)
vmovaps %zmm0, 0x15c0(%rsp)
vmovaps 0x1600(%rsp), %zmm0
vmovaps 0x15c0(%rsp), %zmm1
vaddps %zmm1, %zmm0, %zmm0
vmovaps %zmm0, 0x680(%rsp)
vmovaps 0x7c8240(%rip), %zmm0 # 0x1e1de00
vmovaps %zmm0, 0x480(%rsp)
vmovaps 0x480(%rsp), %zmm2
vmovaps 0x4c0(%rsp), %zmm1
vmovaps 0x7c825e(%rip), %zmm0 # 0x1e1de40
vmovaps %zmm2, 0x1000(%rsp)
vmovaps %zmm1, 0xfc0(%rsp)
vmovaps %zmm0, 0xf80(%rsp)
vmovaps 0x1000(%rsp), %zmm1
vmovaps 0xfc0(%rsp), %zmm0
vmovaps 0xf80(%rsp), %zmm2
vfmadd213ps %zmm2, %zmm1, %zmm0 # zmm0 = (zmm1 * zmm0) + zmm2
vmovaps %zmm0, 0x480(%rsp)
vmovaps 0x480(%rsp), %zmm2
vmovaps 0x4c0(%rsp), %zmm1
vmovaps 0x7c8246(%rip), %zmm0 # 0x1e1de80
vmovaps %zmm2, 0xf40(%rsp)
vmovaps %zmm1, 0xf00(%rsp)
vmovaps %zmm0, 0xec0(%rsp)
vmovaps 0xf40(%rsp), %zmm1
vmovaps 0xf00(%rsp), %zmm0
vmovaps 0xec0(%rsp), %zmm2
vfmadd213ps %zmm2, %zmm1, %zmm0 # zmm0 = (zmm1 * zmm0) + zmm2
vmovaps %zmm0, 0x480(%rsp)
vmovaps 0x480(%rsp), %zmm1
vmovaps 0x4c0(%rsp), %zmm0
vmovaps %zmm1, 0xbc0(%rsp)
vmovaps %zmm0, 0xb80(%rsp)
vmovaps 0xbc0(%rsp), %zmm0
vmovaps 0xb80(%rsp), %zmm1
vmulps %zmm1, %zmm0, %zmm0
vmovaps %zmm0, 0x480(%rsp)
vmovaps 0x480(%rsp), %zmm1
vmovaps 0x800(%rsp), %zmm0
vmovaps %zmm1, 0xe80(%rsp)
vmovaps %zmm0, 0xe40(%rsp)
vmovaps %zmm0, 0xe00(%rsp)
vmovaps 0xe80(%rsp), %zmm1
vmovaps 0xe40(%rsp), %zmm0
vmovaps 0xe00(%rsp), %zmm2
vfmadd213ps %zmm2, %zmm1, %zmm0 # zmm0 = (zmm1 * zmm0) + zmm2
vmovaps %zmm0, 0x480(%rsp)
vmovaps 0x540(%rsp), %zmm0
vmovaps %zmm0, 0x700(%rsp)
vmovaps 0x700(%rsp), %zmm1
vmovaps 0x480(%rsp), %zmm0
vmovaps %zmm1, 0x1880(%rsp)
vmovaps %zmm0, 0x1840(%rsp)
vmovdqa64 0x1880(%rsp), %zmm0
vmovdqa64 0x1840(%rsp), %zmm1
vpandd %zmm1, %zmm0, %zmm0
vmovdqa64 %zmm0, 0x440(%rsp)
vmovaps 0x700(%rsp), %zmm1
vmovaps 0x680(%rsp), %zmm0
vmovaps %zmm1, 0x1d80(%rsp)
vmovaps %zmm0, 0x1d40(%rsp)
vmovdqa64 0x1d80(%rsp), %zmm0
vmovdqa64 0x1d40(%rsp), %zmm1
vpandnd %zmm1, %zmm0, %zmm0
vmovdqa64 %zmm0, 0x400(%rsp)
vmovaps 0x480(%rsp), %zmm1
vmovaps 0x440(%rsp), %zmm0
vmovaps %zmm1, 0xb40(%rsp)
vmovaps %zmm0, 0xb00(%rsp)
vmovaps 0xb40(%rsp), %zmm0
vmovaps 0xb00(%rsp), %zmm1
vsubps %zmm1, %zmm0, %zmm0
vmovaps %zmm0, 0x480(%rsp)
vmovaps 0x680(%rsp), %zmm1
vmovaps 0x400(%rsp), %zmm0
vmovaps %zmm1, 0xac0(%rsp)
vmovaps %zmm0, 0xa80(%rsp)
vmovaps 0xac0(%rsp), %zmm0
vmovaps 0xa80(%rsp), %zmm1
vsubps %zmm1, %zmm0, %zmm0
vmovaps %zmm0, 0x680(%rsp)
vmovaps 0x400(%rsp), %zmm1
vmovaps 0x440(%rsp), %zmm0
vmovaps %zmm1, 0x1580(%rsp)
vmovaps %zmm0, 0x1540(%rsp)
vmovaps 0x1580(%rsp), %zmm0
vmovaps 0x1540(%rsp), %zmm1
vaddps %zmm1, %zmm0, %zmm0
vmovaps %zmm0, 0x780(%rsp)
vmovaps 0x680(%rsp), %zmm1
vmovaps 0x480(%rsp), %zmm0
vmovaps %zmm1, 0x1500(%rsp)
vmovaps %zmm0, 0x14c0(%rsp)
vmovaps 0x1500(%rsp), %zmm0
vmovaps 0x14c0(%rsp), %zmm1
vaddps %zmm1, %zmm0, %zmm0
vmovaps %zmm0, 0x740(%rsp)
vmovaps 0x780(%rsp), %zmm1
vmovaps 0x6c0(%rsp), %zmm0
vmovaps %zmm1, 0x1c80(%rsp)
vmovaps %zmm0, 0x1c40(%rsp)
vmovdqa64 0x1c80(%rsp), %zmm0
vmovdqa64 0x1c40(%rsp), %zmm1
vpxord %zmm1, %zmm0, %zmm0
movq 0x7f8(%rsp), %rax
vmovdqa64 %zmm0, (%rax)
vmovaps 0x740(%rsp), %zmm1
vmovaps 0x500(%rsp), %zmm0
vmovaps %zmm1, 0x1c00(%rsp)
vmovaps %zmm0, 0x1bc0(%rsp)
vmovdqa64 0x1c00(%rsp), %zmm0
vmovdqa64 0x1bc0(%rsp), %zmm1
vpxord %zmm1, %zmm0, %zmm0
movq 0x7f0(%rsp), %rax
vmovdqa64 %zmm0, (%rax)
vmovaps 0x100(%rsp), %zmm0
vpxor %xmm1, %xmm1, %xmm1
vmovaps %zmm1, 0x1c0(%rsp)
vmovaps 0x1c0(%rsp), %zmm1
vcmpeqps %zmm1, %zmm0, %k0
kmovw %k0, 0xbe(%rsp)
vmovaps 0x100(%rsp), %zmm1
movw 0xbe(%rsp), %ax
vmovaps 0xc0(%rsp), %zmm0
vmovaps %zmm1, 0x2c0(%rsp)
movw %ax, 0x2be(%rsp)
vmovaps %zmm1, 0x240(%rsp)
vmovaps %zmm0, 0x200(%rsp)
vmovaps 0x240(%rsp), %zmm1
vmovaps 0x200(%rsp), %zmm0
kmovw 0x2be(%rsp), %k1
vmovaps %zmm1, 0x340(%rsp)
vmovaps %zmm0, 0x300(%rsp)
vmovaps 0x340(%rsp), %zmm1
vmovaps 0x300(%rsp), %zmm2
vmovaps 0x2c0(%rsp), %zmm0
vaddps %zmm2, %zmm1, %zmm0 {%k1}
vmovaps %zmm0, 0x100(%rsp)
vmovaps 0x140(%rsp), %zmm1
vmovaps 0x100(%rsp), %zmm0
vmovaps %zmm1, 0x1e80(%rsp)
vmovaps %zmm0, 0x1e40(%rsp)
vmovaps 0x1e80(%rsp), %zmm0
vdivps 0x1e40(%rsp), %zmm0, %zmm0
vmovaps %zmm0, 0x40(%rsp)
vmovaps 0x40(%rsp), %zmm0
movq %rbp, %rsp
popq %rbp
retq
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,298 |
ncnn::UnaryOp_x86_avx512_functor::unary_op_tan::func_pack8(float vector[8] const&) const
|
__m256 func_pack8(const __m256& x) const
{
return tan256_ps(x);
}
|
pushq %rbp
movq %rsp, %rbp
andq $-0x20, %rsp
subq $0x1380, %rsp # imm = 0x1380
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
movq 0x10(%rsp), %rax
vmovaps (%rax), %ymm0
vmovaps %ymm0, 0xe0(%rsp)
movl $0x322bcc77, 0x1bc(%rsp) # imm = 0x322BCC77
vmovss 0x1bc(%rsp), %xmm0
vmovss %xmm0, 0x136c(%rsp)
vmovss %xmm0, 0x1368(%rsp)
vmovss %xmm0, 0x1364(%rsp)
vmovss %xmm0, 0x1360(%rsp)
vmovss %xmm0, 0x135c(%rsp)
vmovss %xmm0, 0x1358(%rsp)
vmovss %xmm0, 0x1354(%rsp)
vmovss %xmm0, 0x1350(%rsp)
vmovss 0x1354(%rsp), %xmm1
vmovss 0x1350(%rsp), %xmm0
vinsertps $0x10, %xmm1, %xmm0, %xmm0 # xmm0 = xmm0[0],xmm1[0],xmm0[2,3]
vmovss 0x1358(%rsp), %xmm1
vinsertps $0x20, %xmm1, %xmm0, %xmm0 # xmm0 = xmm0[0,1],xmm1[0],xmm0[3]
vmovss 0x135c(%rsp), %xmm1
vinsertps $0x30, %xmm1, %xmm0, %xmm0 # xmm0 = xmm0[0,1,2],xmm1[0]
vmovss 0x1364(%rsp), %xmm2
vmovss 0x1360(%rsp), %xmm1
vinsertps $0x10, %xmm2, %xmm1, %xmm1 # xmm1 = xmm1[0],xmm2[0],xmm1[2,3]
vmovss 0x1368(%rsp), %xmm2
vinsertps $0x20, %xmm2, %xmm1, %xmm1 # xmm1 = xmm1[0,1],xmm2[0],xmm1[3]
vmovss 0x136c(%rsp), %xmm2
vinsertps $0x30, %xmm2, %xmm1, %xmm1 # xmm1 = xmm1[0,1,2],xmm2[0]
vmovaps %xmm1, 0x1330(%rsp)
vmovaps %xmm0, 0x1320(%rsp)
vmovaps 0x1320(%rsp), %ymm0
vmovaps %ymm0, 0x80(%rsp)
vmovaps 0xe0(%rsp), %ymm0
vmovaps %ymm0, 0x3c0(%rsp)
leaq 0xc0(%rsp), %rax
movq %rax, 0x3b8(%rsp)
leaq 0xa0(%rsp), %rax
movq %rax, 0x3b0(%rsp)
vpxor %xmm1, %xmm1, %xmm1
vmovaps %ymm1, 0x460(%rsp)
vmovaps 0x460(%rsp), %ymm0
vmovaps %ymm0, 0x340(%rsp)
vmovaps 0x3c0(%rsp), %ymm0
vmovaps %ymm0, 0x320(%rsp)
vmovaps 0x3c0(%rsp), %ymm0
vmovaps %ymm0, 0x960(%rsp)
vbroadcastss 0x7c6c0b(%rip), %ymm0 # 0x1e1cdb0
vmovaps %ymm0, 0x940(%rsp)
vmovdqa 0x960(%rsp), %ymm0
vmovdqa 0x940(%rsp), %ymm2
vpand %ymm2, %ymm0, %ymm0
vmovdqa %ymm0, 0x3c0(%rsp)
vmovaps 0x320(%rsp), %ymm0
vmovaps %ymm0, 0x920(%rsp)
vbroadcastss 0x7b7cd0(%rip), %ymm0 # 0x1e0deb8
vmovaps %ymm0, 0x900(%rsp)
vmovdqa 0x920(%rsp), %ymm0
vmovdqa 0x900(%rsp), %ymm2
vpand %ymm2, %ymm0, %ymm0
vmovdqa %ymm0, 0x320(%rsp)
vmovaps 0x3c0(%rsp), %ymm0
vmovaps %ymm0, 0x5a0(%rsp)
vbroadcastss 0x7c6b8d(%rip), %ymm0 # 0x1e1cdb8
vmovaps %ymm0, 0x580(%rsp)
vmovaps 0x5a0(%rsp), %ymm0
vmovaps 0x580(%rsp), %ymm2
vmulps %ymm2, %ymm0, %ymm0
vmovaps %ymm0, 0x300(%rsp)
vmovaps 0x300(%rsp), %ymm0
vmovaps %ymm0, 0xaa0(%rsp)
vcvttps2dq 0xaa0(%rsp), %ymm0
vmovdqa %ymm0, 0x2c0(%rsp)
vmovdqa 0x2c0(%rsp), %ymm0
vmovdqa %ymm0, 0xae0(%rsp)
vpbroadcastq 0x7c6b4e(%rip), %ymm0 # 0x1e1cde0
vmovdqa %ymm0, 0xac0(%rsp)
vmovdqa 0xae0(%rsp), %ymm2
vmovdqa 0xac0(%rsp), %ymm0
vmovdqa %ymm2, 0xf60(%rsp)
vmovdqa %ymm0, 0xf40(%rsp)
vmovdqa 0xf60(%rsp), %ymm0
vmovdqa 0xf40(%rsp), %ymm2
vpaddd %ymm2, %ymm0, %ymm0
vmovdqa %ymm0, 0x2c0(%rsp)
vmovdqa 0x2c0(%rsp), %ymm0
vmovdqa %ymm0, 0x1140(%rsp)
vpbroadcastq 0x7c6aef(%rip), %ymm0 # 0x1e1cde8
vmovdqa %ymm0, 0x1120(%rsp)
vmovdqa 0x1140(%rsp), %ymm0
vmovdqa 0x1120(%rsp), %ymm2
vpand %ymm2, %ymm0, %ymm0
vmovdqa %ymm0, 0x2c0(%rsp)
vmovdqa 0x2c0(%rsp), %ymm0
vmovdqa %ymm0, 0x1040(%rsp)
vcvtdq2ps 0x1040(%rsp), %ymm0
vmovaps %ymm0, 0x300(%rsp)
vmovdqa 0x2c0(%rsp), %ymm0
vmovdqa %ymm0, 0x2a0(%rsp)
vmovdqa 0x2c0(%rsp), %ymm0
vmovdqa %ymm0, 0x1100(%rsp)
vpbroadcastq 0x7c6a7e(%rip), %ymm0 # 0x1e1cdf0
vmovdqa %ymm0, 0x10e0(%rsp)
vmovdqa 0x1100(%rsp), %ymm2
vmovdqa 0x10e0(%rsp), %ymm3
vpand %ymm3, %ymm2, %ymm2
vmovdqa %ymm2, 0x2e0(%rsp)
vmovdqa 0x2e0(%rsp), %ymm2
vmovdqa %ymm2, 0xb60(%rsp)
movl $0x1d, 0xb5c(%rsp)
vmovdqa 0xb60(%rsp), %ymm2
movl 0xb5c(%rsp), %eax
vmovdqa %ymm2, 0xfa0(%rsp)
movl %eax, 0xf9c(%rsp)
vmovdqa 0xfa0(%rsp), %ymm2
vmovd 0xf9c(%rsp), %xmm3
vpslld %xmm3, %ymm2, %ymm2
vmovdqa %ymm2, 0x2e0(%rsp)
vmovdqa 0x2c0(%rsp), %ymm2
vmovdqa %ymm2, 0x10c0(%rsp)
vpbroadcastq 0x7c69e7(%rip), %ymm2 # 0x1e1cdf8
vmovdqa %ymm2, 0x10a0(%rsp)
vmovdqa 0x10c0(%rsp), %ymm3
vmovdqa 0x10a0(%rsp), %ymm4
vpand %ymm4, %ymm3, %ymm3
vmovdqa %ymm3, 0x2c0(%rsp)
vmovdqa 0x2c0(%rsp), %ymm3
vmovdqa %ymm3, 0x1180(%rsp)
vpxor %xmm3, %xmm3, %xmm3
vmovdqa %ymm3, 0x1160(%rsp)
vmovdqa 0x1180(%rsp), %ymm3
vmovdqa 0x1160(%rsp), %ymm4
vpcmpeqd %ymm4, %ymm3, %ymm3
vmovdqa %ymm3, 0x2c0(%rsp)
vmovdqa 0x2e0(%rsp), %ymm3
vmovdqa %ymm3, 0xbc0(%rsp)
vmovdqa 0xbc0(%rsp), %ymm3
vmovdqa %ymm3, 0x280(%rsp)
vmovdqa 0x2c0(%rsp), %ymm3
vmovdqa %ymm3, 0xba0(%rsp)
vmovdqa 0xba0(%rsp), %ymm3
vmovdqa %ymm3, 0x260(%rsp)
vbroadcastss 0x7c68f4(%rip), %ymm3 # 0x1e1cdbc
vmovaps %ymm3, 0x380(%rsp)
vbroadcastss 0x7c68e6(%rip), %ymm3 # 0x1e1cdc0
vmovaps %ymm3, 0x360(%rsp)
vbroadcastss 0x7c68d8(%rip), %ymm3 # 0x1e1cdc4
vmovaps %ymm3, 0x340(%rsp)
vmovaps 0x300(%rsp), %ymm5
vmovaps 0x380(%rsp), %ymm4
vmovaps 0x3c0(%rsp), %ymm3
vmovaps %ymm5, 0x8a0(%rsp)
vmovaps %ymm4, 0x880(%rsp)
vmovaps %ymm3, 0x860(%rsp)
vmovaps 0x8a0(%rsp), %ymm5
vmovaps 0x880(%rsp), %ymm4
vmovaps 0x860(%rsp), %ymm3
vmovaps %ymm5, 0xc20(%rsp)
vmovaps %ymm4, 0xc00(%rsp)
vmovaps %ymm3, 0xbe0(%rsp)
vmovaps 0xc20(%rsp), %ymm4
vmovaps 0xc00(%rsp), %ymm3
vmovaps 0xbe0(%rsp), %ymm5
vfmadd213ps %ymm5, %ymm4, %ymm3 # ymm3 = (ymm4 * ymm3) + ymm5
vmovaps %ymm3, 0x3c0(%rsp)
vmovaps 0x300(%rsp), %ymm5
vmovaps 0x360(%rsp), %ymm4
vmovaps 0x3c0(%rsp), %ymm3
vmovaps %ymm5, 0x840(%rsp)
vmovaps %ymm4, 0x820(%rsp)
vmovaps %ymm3, 0x800(%rsp)
vmovaps 0x840(%rsp), %ymm5
vmovaps 0x820(%rsp), %ymm4
vmovaps 0x800(%rsp), %ymm3
vmovaps %ymm5, 0xc80(%rsp)
vmovaps %ymm4, 0xc60(%rsp)
vmovaps %ymm3, 0xc40(%rsp)
vmovaps 0xc80(%rsp), %ymm4
vmovaps 0xc60(%rsp), %ymm3
vmovaps 0xc40(%rsp), %ymm5
vfmadd213ps %ymm5, %ymm4, %ymm3 # ymm3 = (ymm4 * ymm3) + ymm5
vmovaps %ymm3, 0x3c0(%rsp)
vmovaps 0x300(%rsp), %ymm5
vmovaps 0x340(%rsp), %ymm4
vmovaps 0x3c0(%rsp), %ymm3
vmovaps %ymm5, 0x7e0(%rsp)
vmovaps %ymm4, 0x7c0(%rsp)
vmovaps %ymm3, 0x7a0(%rsp)
vmovaps 0x7e0(%rsp), %ymm5
vmovaps 0x7c0(%rsp), %ymm4
vmovaps 0x7a0(%rsp), %ymm3
vmovaps %ymm5, 0xce0(%rsp)
vmovaps %ymm4, 0xcc0(%rsp)
vmovaps %ymm3, 0xca0(%rsp)
vmovaps 0xce0(%rsp), %ymm4
vmovaps 0xcc0(%rsp), %ymm3
vmovaps 0xca0(%rsp), %ymm5
vfmadd213ps %ymm5, %ymm4, %ymm3 # ymm3 = (ymm4 * ymm3) + ymm5
vmovaps %ymm3, 0x3c0(%rsp)
vmovdqa 0x2a0(%rsp), %ymm3
vmovdqa %ymm3, 0x1020(%rsp)
vmovdqa %ymm2, 0x1000(%rsp)
vmovdqa 0x1020(%rsp), %ymm3
vmovdqa 0x1000(%rsp), %ymm2
vmovdqa %ymm3, 0x1080(%rsp)
vmovdqa %ymm2, 0x1060(%rsp)
vmovdqa 0x1080(%rsp), %ymm2
vmovdqa 0x1060(%rsp), %ymm3
vpsubd %ymm3, %ymm2, %ymm2
vmovdqa %ymm2, 0x2a0(%rsp)
vmovdqa 0x2a0(%rsp), %ymm2
vmovdqa %ymm2, 0x12c0(%rsp)
vmovdqa %ymm0, 0x12a0(%rsp)
vmovdqa 0x12c0(%rsp), %ymm0
vmovdqa 0x12a0(%rsp), %ymm2
vpandn %ymm2, %ymm0, %ymm0
vmovdqa %ymm0, 0x2a0(%rsp)
vmovdqa 0x2a0(%rsp), %ymm0
vmovdqa %ymm0, 0xb20(%rsp)
movl $0x1d, 0xb1c(%rsp)
vmovdqa 0xb20(%rsp), %ymm0
movl 0xb1c(%rsp), %eax
vmovdqa %ymm0, 0xfe0(%rsp)
movl %eax, 0xfdc(%rsp)
vmovdqa 0xfe0(%rsp), %ymm0
vmovd 0xfdc(%rsp), %xmm2
vpslld %xmm2, %ymm0, %ymm0
vmovdqa %ymm0, 0x2a0(%rsp)
vmovdqa 0x2a0(%rsp), %ymm0
vmovdqa %ymm0, 0xb80(%rsp)
vmovdqa 0xb80(%rsp), %ymm0
vmovdqa %ymm0, 0x240(%rsp)
vmovaps 0x320(%rsp), %ymm2
vmovaps 0x280(%rsp), %ymm0
vmovaps %ymm2, 0x1240(%rsp)
vmovaps %ymm0, 0x1220(%rsp)
vmovdqa 0x1240(%rsp), %ymm0
vmovdqa 0x1220(%rsp), %ymm2
vpxor %ymm2, %ymm0, %ymm0
vmovdqa %ymm0, 0x320(%rsp)
vmovaps 0x3c0(%rsp), %ymm0
vmovaps %ymm0, 0x560(%rsp)
vmovaps %ymm0, 0x540(%rsp)
vmovaps 0x560(%rsp), %ymm0
vmovaps 0x540(%rsp), %ymm2
vmulps %ymm2, %ymm0, %ymm0
vmovaps %ymm0, 0x220(%rsp)
vbroadcastss 0x7c6576(%rip), %ymm0 # 0x1e1cdc8
vmovaps %ymm0, 0x300(%rsp)
vmovaps 0x300(%rsp), %ymm2
vmovaps 0x220(%rsp), %ymm0
vmovaps %ymm2, 0x780(%rsp)
vmovaps %ymm0, 0x760(%rsp)
vbroadcastss 0x7c6544(%rip), %ymm0 # 0x1e1cdcc
vmovaps %ymm0, 0x740(%rsp)
vmovaps 0x780(%rsp), %ymm3
vmovaps 0x760(%rsp), %ymm2
vmovaps 0x740(%rsp), %ymm0
vmovaps %ymm3, 0xd40(%rsp)
vmovaps %ymm2, 0xd20(%rsp)
vmovaps %ymm0, 0xd00(%rsp)
vmovaps 0xd40(%rsp), %ymm2
vmovaps 0xd20(%rsp), %ymm0
vmovaps 0xd00(%rsp), %ymm3
vfmadd213ps %ymm3, %ymm2, %ymm0 # ymm0 = (ymm2 * ymm0) + ymm3
vmovaps %ymm0, 0x300(%rsp)
vmovaps 0x300(%rsp), %ymm2
vmovaps 0x220(%rsp), %ymm0
vmovaps %ymm2, 0x720(%rsp)
vmovaps %ymm0, 0x700(%rsp)
vbroadcastss 0x7c64b3(%rip), %ymm0 # 0x1e1cdd0
vmovaps %ymm0, 0x6e0(%rsp)
vmovaps 0x720(%rsp), %ymm3
vmovaps 0x700(%rsp), %ymm2
vmovaps 0x6e0(%rsp), %ymm0
vmovaps %ymm3, 0xda0(%rsp)
vmovaps %ymm2, 0xd80(%rsp)
vmovaps %ymm0, 0xd60(%rsp)
vmovaps 0xda0(%rsp), %ymm2
vmovaps 0xd80(%rsp), %ymm0
vmovaps 0xd60(%rsp), %ymm3
vfmadd213ps %ymm3, %ymm2, %ymm0 # ymm0 = (ymm2 * ymm0) + ymm3
vmovaps %ymm0, 0x300(%rsp)
vmovaps 0x300(%rsp), %ymm2
vmovaps 0x220(%rsp), %ymm0
vmovaps %ymm2, 0x520(%rsp)
vmovaps %ymm0, 0x500(%rsp)
vmovaps 0x520(%rsp), %ymm0
vmovaps 0x500(%rsp), %ymm2
vmulps %ymm2, %ymm0, %ymm0
vmovaps %ymm0, 0x300(%rsp)
vmovaps 0x300(%rsp), %ymm2
vmovaps 0x220(%rsp), %ymm0
vmovaps %ymm2, 0x4e0(%rsp)
vmovaps %ymm0, 0x4c0(%rsp)
vmovaps 0x4e0(%rsp), %ymm0
vmovaps 0x4c0(%rsp), %ymm2
vmulps %ymm2, %ymm0, %ymm0
vmovaps %ymm0, 0x300(%rsp)
vmovaps 0x220(%rsp), %ymm2
vmovaps 0x300(%rsp), %ymm0
vmovaps %ymm2, 0x9c0(%rsp)
vbroadcastss 0x7b25ed(%rip), %ymm2 # 0x1e0901c
vmovaps %ymm2, 0x9a0(%rsp)
vmovaps %ymm0, 0x980(%rsp)
vmovaps 0x9c0(%rsp), %ymm3
vmovaps 0x9a0(%rsp), %ymm2
vmovaps 0x980(%rsp), %ymm0
vmovaps %ymm3, 0xf20(%rsp)
vmovaps %ymm2, 0xf00(%rsp)
vmovaps %ymm0, 0xee0(%rsp)
vmovaps 0xf20(%rsp), %ymm2
vmovaps 0xf00(%rsp), %ymm0
vmovaps 0xee0(%rsp), %ymm3
vfnmadd213ps %ymm3, %ymm2, %ymm0 # ymm0 = -(ymm2 * ymm0) + ymm3
vmovaps %ymm0, 0x300(%rsp)
vmovaps 0x300(%rsp), %ymm0
vmovaps %ymm0, 0xa80(%rsp)
vbroadcastss 0x7b2549(%rip), %ymm0 # 0x1e09004
vmovaps %ymm0, 0xa60(%rsp)
vmovaps 0xa80(%rsp), %ymm0
vmovaps 0xa60(%rsp), %ymm2
vaddps %ymm2, %ymm0, %ymm0
vmovaps %ymm0, 0x300(%rsp)
vbroadcastss 0x7c62e8(%rip), %ymm0 # 0x1e1cdd4
vmovaps %ymm0, 0x200(%rsp)
vmovaps 0x200(%rsp), %ymm2
vmovaps 0x220(%rsp), %ymm0
vmovaps %ymm2, 0x6c0(%rsp)
vmovaps %ymm0, 0x6a0(%rsp)
vbroadcastss 0x7c62b6(%rip), %ymm0 # 0x1e1cdd8
vmovaps %ymm0, 0x680(%rsp)
vmovaps 0x6c0(%rsp), %ymm3
vmovaps 0x6a0(%rsp), %ymm2
vmovaps 0x680(%rsp), %ymm0
vmovaps %ymm3, 0xe00(%rsp)
vmovaps %ymm2, 0xde0(%rsp)
vmovaps %ymm0, 0xdc0(%rsp)
vmovaps 0xe00(%rsp), %ymm2
vmovaps 0xde0(%rsp), %ymm0
vmovaps 0xdc0(%rsp), %ymm3
vfmadd213ps %ymm3, %ymm2, %ymm0 # ymm0 = (ymm2 * ymm0) + ymm3
vmovaps %ymm0, 0x200(%rsp)
vmovaps 0x200(%rsp), %ymm2
vmovaps 0x220(%rsp), %ymm0
vmovaps %ymm2, 0x660(%rsp)
vmovaps %ymm0, 0x640(%rsp)
vbroadcastss 0x7c6225(%rip), %ymm0 # 0x1e1cddc
vmovaps %ymm0, 0x620(%rsp)
vmovaps 0x660(%rsp), %ymm3
vmovaps 0x640(%rsp), %ymm2
vmovaps 0x620(%rsp), %ymm0
vmovaps %ymm3, 0xe60(%rsp)
vmovaps %ymm2, 0xe40(%rsp)
vmovaps %ymm0, 0xe20(%rsp)
vmovaps 0xe60(%rsp), %ymm2
vmovaps 0xe40(%rsp), %ymm0
vmovaps 0xe20(%rsp), %ymm3
vfmadd213ps %ymm3, %ymm2, %ymm0 # ymm0 = (ymm2 * ymm0) + ymm3
vmovaps %ymm0, 0x200(%rsp)
vmovaps 0x200(%rsp), %ymm2
vmovaps 0x220(%rsp), %ymm0
vmovaps %ymm2, 0x4a0(%rsp)
vmovaps %ymm0, 0x480(%rsp)
vmovaps 0x4a0(%rsp), %ymm0
vmovaps 0x480(%rsp), %ymm2
vmulps %ymm2, %ymm0, %ymm0
vmovaps %ymm0, 0x200(%rsp)
vmovaps 0x200(%rsp), %ymm2
vmovaps 0x3c0(%rsp), %ymm0
vmovaps %ymm2, 0x600(%rsp)
vmovaps %ymm0, 0x5e0(%rsp)
vmovaps %ymm0, 0x5c0(%rsp)
vmovaps 0x600(%rsp), %ymm3
vmovaps 0x5e0(%rsp), %ymm2
vmovaps 0x5c0(%rsp), %ymm0
vmovaps %ymm3, 0xec0(%rsp)
vmovaps %ymm2, 0xea0(%rsp)
vmovaps %ymm0, 0xe80(%rsp)
vmovaps 0xec0(%rsp), %ymm2
vmovaps 0xea0(%rsp), %ymm0
vmovaps 0xe80(%rsp), %ymm3
vfmadd213ps %ymm3, %ymm2, %ymm0 # ymm0 = (ymm2 * ymm0) + ymm3
vmovaps %ymm0, 0x200(%rsp)
vmovaps 0x260(%rsp), %ymm0
vmovaps %ymm0, 0x340(%rsp)
vmovaps 0x340(%rsp), %ymm2
vmovaps 0x200(%rsp), %ymm0
vmovaps %ymm2, 0x8e0(%rsp)
vmovaps %ymm0, 0x8c0(%rsp)
vmovdqa 0x8e0(%rsp), %ymm0
vmovdqa 0x8c0(%rsp), %ymm2
vpand %ymm2, %ymm0, %ymm0
vmovdqa %ymm0, 0x1e0(%rsp)
vmovaps 0x340(%rsp), %ymm2
vmovaps 0x300(%rsp), %ymm0
vmovaps %ymm2, 0x1280(%rsp)
vmovaps %ymm0, 0x1260(%rsp)
vmovdqa 0x1280(%rsp), %ymm0
vmovdqa 0x1260(%rsp), %ymm2
vpandn %ymm2, %ymm0, %ymm0
vmovdqa %ymm0, 0x1c0(%rsp)
vmovaps 0x200(%rsp), %ymm2
vmovaps 0x1e0(%rsp), %ymm0
vmovaps %ymm2, 0x440(%rsp)
vmovaps %ymm0, 0x420(%rsp)
vmovaps 0x440(%rsp), %ymm0
vmovaps 0x420(%rsp), %ymm2
vsubps %ymm2, %ymm0, %ymm0
vmovaps %ymm0, 0x200(%rsp)
vmovaps 0x300(%rsp), %ymm2
vmovaps 0x1c0(%rsp), %ymm0
vmovaps %ymm2, 0x400(%rsp)
vmovaps %ymm0, 0x3e0(%rsp)
vmovaps 0x400(%rsp), %ymm0
vmovaps 0x3e0(%rsp), %ymm2
vsubps %ymm2, %ymm0, %ymm0
vmovaps %ymm0, 0x300(%rsp)
vmovaps 0x1c0(%rsp), %ymm2
vmovaps 0x1e0(%rsp), %ymm0
vmovaps %ymm2, 0xa40(%rsp)
vmovaps %ymm0, 0xa20(%rsp)
vmovaps 0xa40(%rsp), %ymm0
vmovaps 0xa20(%rsp), %ymm2
vaddps %ymm2, %ymm0, %ymm0
vmovaps %ymm0, 0x380(%rsp)
vmovaps 0x300(%rsp), %ymm2
vmovaps 0x200(%rsp), %ymm0
vmovaps %ymm2, 0xa00(%rsp)
vmovaps %ymm0, 0x9e0(%rsp)
vmovaps 0xa00(%rsp), %ymm0
vmovaps 0x9e0(%rsp), %ymm2
vaddps %ymm2, %ymm0, %ymm0
vmovaps %ymm0, 0x360(%rsp)
vmovaps 0x380(%rsp), %ymm2
vmovaps 0x320(%rsp), %ymm0
vmovaps %ymm2, 0x1200(%rsp)
vmovaps %ymm0, 0x11e0(%rsp)
vmovdqa 0x1200(%rsp), %ymm0
vmovdqa 0x11e0(%rsp), %ymm2
vpxor %ymm2, %ymm0, %ymm0
movq 0x3b8(%rsp), %rax
vmovdqa %ymm0, (%rax)
vmovaps 0x360(%rsp), %ymm2
vmovaps 0x240(%rsp), %ymm0
vmovaps %ymm2, 0x11c0(%rsp)
vmovaps %ymm0, 0x11a0(%rsp)
vmovdqa 0x11c0(%rsp), %ymm0
vmovdqa 0x11a0(%rsp), %ymm2
vpxor %ymm2, %ymm0, %ymm0
movq 0x3b0(%rsp), %rax
vmovdqa %ymm0, (%rax)
vmovaps 0xa0(%rsp), %ymm0
vmovaps %ymm1, 0x100(%rsp)
vmovaps 0x100(%rsp), %ymm1
vcmpeqps %ymm1, %ymm0, %k0
vpmovm2d %k0, %ymm0
vmovaps %ymm0, 0x60(%rsp)
vmovaps 0x80(%rsp), %ymm1
vmovaps 0x60(%rsp), %ymm0
vmovaps %ymm1, 0x140(%rsp)
vmovaps %ymm0, 0x120(%rsp)
vmovaps 0x140(%rsp), %ymm0
vmovaps 0x120(%rsp), %ymm1
vpand %ymm1, %ymm0, %ymm0
vmovaps %ymm0, 0x40(%rsp)
vmovaps 0xa0(%rsp), %ymm1
vmovaps 0x40(%rsp), %ymm0
vmovaps %ymm1, 0x180(%rsp)
vmovaps %ymm0, 0x160(%rsp)
vmovaps 0x180(%rsp), %ymm0
vaddps 0x160(%rsp), %ymm0, %ymm0
vmovaps %ymm0, 0xa0(%rsp)
vmovaps 0xc0(%rsp), %ymm1
vmovaps 0xa0(%rsp), %ymm0
vmovaps %ymm1, 0x1300(%rsp)
vmovaps %ymm0, 0x12e0(%rsp)
vmovaps 0x1300(%rsp), %ymm0
vdivps 0x12e0(%rsp), %ymm0, %ymm0
vmovaps %ymm0, 0x20(%rsp)
vmovaps 0x20(%rsp), %ymm0
movq %rbp, %rsp
popq %rbp
retq
nopl (%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
2,113,299 |
ncnn::UnaryOp_x86_avx512_functor::unary_op_tan::func_pack4(float vector[4] const&) const
|
__m128 func_pack4(const __m128& x) const
{
return tan_ps(x);
}
|
subq $0x8d8, %rsp # imm = 0x8D8
movq %rdi, -0x78(%rsp)
movq %rsi, -0x80(%rsp)
movq -0x80(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm0, -0x10(%rsp)
movl $0x322bcc77, 0x1c(%rsp) # imm = 0x322BCC77
vbroadcastss 0x1c(%rsp), %xmm0
vmovaps %xmm0, (%rsp)
vmovaps (%rsp), %xmm0
vmovaps %xmm0, -0x40(%rsp)
vmovaps -0x10(%rsp), %xmm0
vmovaps %xmm0, 0x170(%rsp)
leaq -0x20(%rsp), %rax
movq %rax, 0x168(%rsp)
leaq -0x30(%rsp), %rax
movq %rax, 0x160(%rsp)
vpxor %xmm0, %xmm0, %xmm0
vmovaps %xmm0, 0x1e0(%rsp)
vmovaps 0x1e0(%rsp), %xmm1
vmovaps %xmm1, 0x130(%rsp)
vmovaps 0x170(%rsp), %xmm1
vmovaps %xmm1, 0x120(%rsp)
vmovaps 0x170(%rsp), %xmm1
vmovaps %xmm1, 0x360(%rsp)
vbroadcastss 0x7c5cec(%rip), %xmm1 # 0x1e1cdb0
vmovaps %xmm1, 0x350(%rsp)
vmovdqa 0x360(%rsp), %xmm1
vmovdqa 0x350(%rsp), %xmm2
vpand %xmm2, %xmm1, %xmm1
vmovdqa %xmm1, 0x170(%rsp)
vmovaps 0x120(%rsp), %xmm1
vmovaps %xmm1, 0x340(%rsp)
vbroadcastss 0x7b6db1(%rip), %xmm1 # 0x1e0deb8
vmovaps %xmm1, 0x330(%rsp)
vmovdqa 0x340(%rsp), %xmm1
vmovdqa 0x330(%rsp), %xmm2
vpand %xmm2, %xmm1, %xmm1
vmovdqa %xmm1, 0x120(%rsp)
vmovaps 0x170(%rsp), %xmm1
vmovaps %xmm1, 0x280(%rsp)
vbroadcastss 0x7c5c6e(%rip), %xmm1 # 0x1e1cdb8
vmovaps %xmm1, 0x270(%rsp)
vmovaps 0x280(%rsp), %xmm1
vmovaps 0x270(%rsp), %xmm2
vmulps %xmm2, %xmm1, %xmm1
vmovaps %xmm1, 0x110(%rsp)
vmovaps 0x110(%rsp), %xmm1
vmovaps %xmm1, 0x2f0(%rsp)
vcvttps2dq 0x2f0(%rsp), %xmm1
vmovdqa %xmm1, 0xf0(%rsp)
vmovdqa 0xf0(%rsp), %xmm1
vmovdqa %xmm1, 0x530(%rsp)
vpbroadcastq 0x7c5c2f(%rip), %xmm1 # 0x1e1cde0
vmovdqa %xmm1, 0x520(%rsp)
vmovdqa 0x530(%rsp), %xmm1
vmovdqa 0x520(%rsp), %xmm2
vpaddd %xmm2, %xmm1, %xmm1
vmovdqa %xmm1, 0xf0(%rsp)
vmovdqa 0xf0(%rsp), %xmm1
vmovdqa %xmm1, 0x7d0(%rsp)
vpbroadcastq 0x7c5bf4(%rip), %xmm1 # 0x1e1cde8
vmovdqa %xmm1, 0x7c0(%rsp)
vmovdqa 0x7d0(%rsp), %xmm1
vmovdqa 0x7c0(%rsp), %xmm2
vpand %xmm2, %xmm1, %xmm1
vmovdqa %xmm1, 0xf0(%rsp)
vmovdqa 0xf0(%rsp), %xmm1
vmovdqa %xmm1, 0x300(%rsp)
vcvtdq2ps 0x300(%rsp), %xmm1
vmovaps %xmm1, 0x110(%rsp)
vmovdqa 0xf0(%rsp), %xmm1
vmovdqa %xmm1, 0xe0(%rsp)
vmovdqa 0xf0(%rsp), %xmm1
vmovdqa %xmm1, 0x7b0(%rsp)
vpbroadcastq 0x7c5b83(%rip), %xmm1 # 0x1e1cdf0
vmovdqa %xmm1, 0x7a0(%rsp)
vmovdqa 0x7b0(%rsp), %xmm2
vmovdqa 0x7a0(%rsp), %xmm3
vpand %xmm3, %xmm2, %xmm2
vmovdqa %xmm2, 0x100(%rsp)
vmovdqa 0x100(%rsp), %xmm2
vmovdqa %xmm2, 0x570(%rsp)
movl $0x1d, 0x56c(%rsp)
vmovdqa 0x570(%rsp), %xmm2
vmovd 0x56c(%rsp), %xmm3
vpslld %xmm3, %xmm2, %xmm2
vmovdqa %xmm2, 0x100(%rsp)
vmovdqa 0x100(%rsp), %xmm2
vmovdqa %xmm2, 0x5a0(%rsp)
vmovdqa 0x5a0(%rsp), %xmm2
vmovdqa %xmm2, 0xd0(%rsp)
vmovdqa 0xf0(%rsp), %xmm2
vmovdqa %xmm2, 0x790(%rsp)
vpbroadcastq 0x7c5ae8(%rip), %xmm2 # 0x1e1cdf8
vmovdqa %xmm2, 0x780(%rsp)
vmovdqa 0x790(%rsp), %xmm3
vmovdqa 0x780(%rsp), %xmm4
vpand %xmm4, %xmm3, %xmm3
vmovdqa %xmm3, 0xf0(%rsp)
vmovdqa 0xf0(%rsp), %xmm4
vpxor %xmm3, %xmm3, %xmm3
vmovdqa %xmm3, 0x800(%rsp)
vmovdqa 0x800(%rsp), %xmm3
vmovdqa %xmm4, 0x7f0(%rsp)
vmovdqa %xmm3, 0x7e0(%rsp)
vmovdqa 0x7f0(%rsp), %xmm3
vmovdqa 0x7e0(%rsp), %xmm4
vpcmpeqd %xmm4, %xmm3, %xmm3
vmovdqa %xmm3, 0xf0(%rsp)
vmovdqa 0xf0(%rsp), %xmm3
vmovdqa %xmm3, 0x590(%rsp)
vmovdqa 0x590(%rsp), %xmm3
vmovdqa %xmm3, 0xc0(%rsp)
vbroadcastss 0x7c5a07(%rip), %xmm3 # 0x1e1cdbc
vmovaps %xmm3, 0x150(%rsp)
vbroadcastss 0x7c59f9(%rip), %xmm3 # 0x1e1cdc0
vmovaps %xmm3, 0x140(%rsp)
vbroadcastss 0x7c59eb(%rip), %xmm3 # 0x1e1cdc4
vmovaps %xmm3, 0x130(%rsp)
vmovaps 0x110(%rsp), %xmm5
vmovaps 0x150(%rsp), %xmm4
vmovaps 0x170(%rsp), %xmm3
vmovaps %xmm5, 0x510(%rsp)
vmovaps %xmm4, 0x500(%rsp)
vmovaps %xmm3, 0x4f0(%rsp)
vmovaps 0x510(%rsp), %xmm5
vmovaps 0x500(%rsp), %xmm4
vmovaps 0x4f0(%rsp), %xmm3
vmovaps %xmm5, 0x600(%rsp)
vmovaps %xmm4, 0x5f0(%rsp)
vmovaps %xmm3, 0x5e0(%rsp)
vmovaps 0x600(%rsp), %xmm4
vmovaps 0x5f0(%rsp), %xmm3
vmovaps 0x5e0(%rsp), %xmm5
vfmadd213ps %xmm5, %xmm4, %xmm3 # xmm3 = (xmm4 * xmm3) + xmm5
vmovaps %xmm3, 0x170(%rsp)
vmovaps 0x110(%rsp), %xmm5
vmovaps 0x140(%rsp), %xmm4
vmovaps 0x170(%rsp), %xmm3
vmovaps %xmm5, 0x4e0(%rsp)
vmovaps %xmm4, 0x4d0(%rsp)
vmovaps %xmm3, 0x4c0(%rsp)
vmovaps 0x4e0(%rsp), %xmm5
vmovaps 0x4d0(%rsp), %xmm4
vmovaps 0x4c0(%rsp), %xmm3
vmovaps %xmm5, 0x630(%rsp)
vmovaps %xmm4, 0x620(%rsp)
vmovaps %xmm3, 0x610(%rsp)
vmovaps 0x630(%rsp), %xmm4
vmovaps 0x620(%rsp), %xmm3
vmovaps 0x610(%rsp), %xmm5
vfmadd213ps %xmm5, %xmm4, %xmm3 # xmm3 = (xmm4 * xmm3) + xmm5
vmovaps %xmm3, 0x170(%rsp)
vmovaps 0x110(%rsp), %xmm5
vmovaps 0x130(%rsp), %xmm4
vmovaps 0x170(%rsp), %xmm3
vmovaps %xmm5, 0x4b0(%rsp)
vmovaps %xmm4, 0x4a0(%rsp)
vmovaps %xmm3, 0x490(%rsp)
vmovaps 0x4b0(%rsp), %xmm5
vmovaps 0x4a0(%rsp), %xmm4
vmovaps 0x490(%rsp), %xmm3
vmovaps %xmm5, 0x660(%rsp)
vmovaps %xmm4, 0x650(%rsp)
vmovaps %xmm3, 0x640(%rsp)
vmovaps 0x660(%rsp), %xmm4
vmovaps 0x650(%rsp), %xmm3
vmovaps 0x640(%rsp), %xmm5
vfmadd213ps %xmm5, %xmm4, %xmm3 # xmm3 = (xmm4 * xmm3) + xmm5
vmovaps %xmm3, 0x170(%rsp)
vmovdqa 0xe0(%rsp), %xmm3
vmovdqa %xmm3, 0x770(%rsp)
vmovdqa %xmm2, 0x760(%rsp)
vmovdqa 0x770(%rsp), %xmm2
vmovdqa 0x760(%rsp), %xmm3
vpsubd %xmm3, %xmm2, %xmm2
vmovdqa %xmm2, 0xe0(%rsp)
vmovdqa 0xe0(%rsp), %xmm2
vmovdqa %xmm2, 0x880(%rsp)
vmovdqa %xmm1, 0x870(%rsp)
vmovdqa 0x880(%rsp), %xmm1
vmovdqa 0x870(%rsp), %xmm2
vpandn %xmm2, %xmm1, %xmm1
vmovdqa %xmm1, 0xe0(%rsp)
vmovdqa 0xe0(%rsp), %xmm1
vmovdqa %xmm1, 0x550(%rsp)
movl $0x1d, 0x54c(%rsp)
vmovdqa 0x550(%rsp), %xmm1
vmovd 0x54c(%rsp), %xmm2
vpslld %xmm2, %xmm1, %xmm1
vmovdqa %xmm1, 0xe0(%rsp)
vmovdqa 0xe0(%rsp), %xmm1
vmovdqa %xmm1, 0x580(%rsp)
vmovdqa 0x580(%rsp), %xmm1
vmovdqa %xmm1, 0xb0(%rsp)
vmovaps 0x120(%rsp), %xmm2
vmovaps 0xd0(%rsp), %xmm1
vmovaps %xmm2, 0x860(%rsp)
vmovaps %xmm1, 0x850(%rsp)
vmovdqa 0x860(%rsp), %xmm1
vmovdqa 0x850(%rsp), %xmm2
vpxor %xmm2, %xmm1, %xmm1
vmovdqa %xmm1, 0x120(%rsp)
vmovaps 0x170(%rsp), %xmm1
vmovaps %xmm1, 0x260(%rsp)
vmovaps %xmm1, 0x250(%rsp)
vmovaps 0x260(%rsp), %xmm1
vmovaps 0x250(%rsp), %xmm2
vmulps %xmm2, %xmm1, %xmm1
vmovaps %xmm1, 0xa0(%rsp)
vbroadcastss 0x7c56cd(%rip), %xmm1 # 0x1e1cdc8
vmovaps %xmm1, 0x110(%rsp)
vmovaps 0x110(%rsp), %xmm2
vmovaps 0xa0(%rsp), %xmm1
vmovaps %xmm2, 0x480(%rsp)
vmovaps %xmm1, 0x470(%rsp)
vbroadcastss 0x7c569b(%rip), %xmm1 # 0x1e1cdcc
vmovaps %xmm1, 0x460(%rsp)
vmovaps 0x480(%rsp), %xmm3
vmovaps 0x470(%rsp), %xmm2
vmovaps 0x460(%rsp), %xmm1
vmovaps %xmm3, 0x690(%rsp)
vmovaps %xmm2, 0x680(%rsp)
vmovaps %xmm1, 0x670(%rsp)
vmovaps 0x690(%rsp), %xmm2
vmovaps 0x680(%rsp), %xmm1
vmovaps 0x670(%rsp), %xmm3
vfmadd213ps %xmm3, %xmm2, %xmm1 # xmm1 = (xmm2 * xmm1) + xmm3
vmovaps %xmm1, 0x110(%rsp)
vmovaps 0x110(%rsp), %xmm2
vmovaps 0xa0(%rsp), %xmm1
vmovaps %xmm2, 0x450(%rsp)
vmovaps %xmm1, 0x440(%rsp)
vbroadcastss 0x7c560a(%rip), %xmm1 # 0x1e1cdd0
vmovaps %xmm1, 0x430(%rsp)
vmovaps 0x450(%rsp), %xmm3
vmovaps 0x440(%rsp), %xmm2
vmovaps 0x430(%rsp), %xmm1
vmovaps %xmm3, 0x6c0(%rsp)
vmovaps %xmm2, 0x6b0(%rsp)
vmovaps %xmm1, 0x6a0(%rsp)
vmovaps 0x6c0(%rsp), %xmm2
vmovaps 0x6b0(%rsp), %xmm1
vmovaps 0x6a0(%rsp), %xmm3
vfmadd213ps %xmm3, %xmm2, %xmm1 # xmm1 = (xmm2 * xmm1) + xmm3
vmovaps %xmm1, 0x110(%rsp)
vmovaps 0x110(%rsp), %xmm2
vmovaps 0xa0(%rsp), %xmm1
vmovaps %xmm2, 0x240(%rsp)
vmovaps %xmm1, 0x230(%rsp)
vmovaps 0x240(%rsp), %xmm1
vmovaps 0x230(%rsp), %xmm2
vmulps %xmm2, %xmm1, %xmm1
vmovaps %xmm1, 0x110(%rsp)
vmovaps 0x110(%rsp), %xmm2
vmovaps 0xa0(%rsp), %xmm1
vmovaps %xmm2, 0x220(%rsp)
vmovaps %xmm1, 0x210(%rsp)
vmovaps 0x220(%rsp), %xmm1
vmovaps 0x210(%rsp), %xmm2
vmulps %xmm2, %xmm1, %xmm1
vmovaps %xmm1, 0x110(%rsp)
vmovaps 0xa0(%rsp), %xmm2
vmovaps 0x110(%rsp), %xmm1
vmovaps %xmm2, 0x390(%rsp)
vbroadcastss 0x7b1744(%rip), %xmm2 # 0x1e0901c
vmovaps %xmm2, 0x380(%rsp)
vmovaps %xmm1, 0x370(%rsp)
vmovaps 0x390(%rsp), %xmm3
vmovaps 0x380(%rsp), %xmm2
vmovaps 0x370(%rsp), %xmm1
vmovaps %xmm3, 0x5d0(%rsp)
vmovaps %xmm2, 0x5c0(%rsp)
vmovaps %xmm1, 0x5b0(%rsp)
vmovaps 0x5d0(%rsp), %xmm2
vmovaps 0x5c0(%rsp), %xmm1
vmovaps 0x5b0(%rsp), %xmm3
vfnmadd213ps %xmm3, %xmm2, %xmm1 # xmm1 = -(xmm2 * xmm1) + xmm3
vmovaps %xmm1, 0x110(%rsp)
vmovaps 0x110(%rsp), %xmm1
vmovaps %xmm1, 0x2e0(%rsp)
vbroadcastss 0x7b16a0(%rip), %xmm1 # 0x1e09004
vmovaps %xmm1, 0x2d0(%rsp)
vmovaps 0x2e0(%rsp), %xmm1
vmovaps 0x2d0(%rsp), %xmm2
vaddps %xmm2, %xmm1, %xmm1
vmovaps %xmm1, 0x110(%rsp)
vbroadcastss 0x7c543f(%rip), %xmm1 # 0x1e1cdd4
vmovaps %xmm1, 0x90(%rsp)
vmovaps 0x90(%rsp), %xmm2
vmovaps 0xa0(%rsp), %xmm1
vmovaps %xmm2, 0x420(%rsp)
vmovaps %xmm1, 0x410(%rsp)
vbroadcastss 0x7c540d(%rip), %xmm1 # 0x1e1cdd8
vmovaps %xmm1, 0x400(%rsp)
vmovaps 0x420(%rsp), %xmm3
vmovaps 0x410(%rsp), %xmm2
vmovaps 0x400(%rsp), %xmm1
vmovaps %xmm3, 0x6f0(%rsp)
vmovaps %xmm2, 0x6e0(%rsp)
vmovaps %xmm1, 0x6d0(%rsp)
vmovaps 0x6f0(%rsp), %xmm2
vmovaps 0x6e0(%rsp), %xmm1
vmovaps 0x6d0(%rsp), %xmm3
vfmadd213ps %xmm3, %xmm2, %xmm1 # xmm1 = (xmm2 * xmm1) + xmm3
vmovaps %xmm1, 0x90(%rsp)
vmovaps 0x90(%rsp), %xmm2
vmovaps 0xa0(%rsp), %xmm1
vmovaps %xmm2, 0x3f0(%rsp)
vmovaps %xmm1, 0x3e0(%rsp)
vbroadcastss 0x7c537c(%rip), %xmm1 # 0x1e1cddc
vmovaps %xmm1, 0x3d0(%rsp)
vmovaps 0x3f0(%rsp), %xmm3
vmovaps 0x3e0(%rsp), %xmm2
vmovaps 0x3d0(%rsp), %xmm1
vmovaps %xmm3, 0x720(%rsp)
vmovaps %xmm2, 0x710(%rsp)
vmovaps %xmm1, 0x700(%rsp)
vmovaps 0x720(%rsp), %xmm2
vmovaps 0x710(%rsp), %xmm1
vmovaps 0x700(%rsp), %xmm3
vfmadd213ps %xmm3, %xmm2, %xmm1 # xmm1 = (xmm2 * xmm1) + xmm3
vmovaps %xmm1, 0x90(%rsp)
vmovaps 0x90(%rsp), %xmm2
vmovaps 0xa0(%rsp), %xmm1
vmovaps %xmm2, 0x200(%rsp)
vmovaps %xmm1, 0x1f0(%rsp)
vmovaps 0x200(%rsp), %xmm1
vmovaps 0x1f0(%rsp), %xmm2
vmulps %xmm2, %xmm1, %xmm1
vmovaps %xmm1, 0x90(%rsp)
vmovaps 0x90(%rsp), %xmm2
vmovaps 0x170(%rsp), %xmm1
vmovaps %xmm2, 0x3c0(%rsp)
vmovaps %xmm1, 0x3b0(%rsp)
vmovaps %xmm1, 0x3a0(%rsp)
vmovaps 0x3c0(%rsp), %xmm3
vmovaps 0x3b0(%rsp), %xmm2
vmovaps 0x3a0(%rsp), %xmm1
vmovaps %xmm3, 0x750(%rsp)
vmovaps %xmm2, 0x740(%rsp)
vmovaps %xmm1, 0x730(%rsp)
vmovaps 0x750(%rsp), %xmm2
vmovaps 0x740(%rsp), %xmm1
vmovaps 0x730(%rsp), %xmm3
vfmadd213ps %xmm3, %xmm2, %xmm1 # xmm1 = (xmm2 * xmm1) + xmm3
vmovaps %xmm1, 0x90(%rsp)
vmovaps 0xc0(%rsp), %xmm1
vmovaps %xmm1, 0x130(%rsp)
vmovaps 0x130(%rsp), %xmm2
vmovaps 0x90(%rsp), %xmm1
vmovaps %xmm2, 0x320(%rsp)
vmovaps %xmm1, 0x310(%rsp)
vmovdqa 0x320(%rsp), %xmm1
vmovdqa 0x310(%rsp), %xmm2
vpand %xmm2, %xmm1, %xmm1
vmovdqa %xmm1, 0x80(%rsp)
vmovaps 0x130(%rsp), %xmm2
vmovaps 0x110(%rsp), %xmm1
vmovaps %xmm2, 0x190(%rsp)
vmovaps %xmm1, 0x180(%rsp)
vmovdqa 0x190(%rsp), %xmm1
vmovdqa 0x180(%rsp), %xmm2
vpandn %xmm2, %xmm1, %xmm1
vmovdqa %xmm1, 0x70(%rsp)
vmovaps 0x90(%rsp), %xmm2
vmovaps 0x80(%rsp), %xmm1
vmovaps %xmm2, 0x1d0(%rsp)
vmovaps %xmm1, 0x1c0(%rsp)
vmovaps 0x1d0(%rsp), %xmm1
vmovaps 0x1c0(%rsp), %xmm2
vsubps %xmm2, %xmm1, %xmm1
vmovaps %xmm1, 0x90(%rsp)
vmovaps 0x110(%rsp), %xmm2
vmovaps 0x70(%rsp), %xmm1
vmovaps %xmm2, 0x1b0(%rsp)
vmovaps %xmm1, 0x1a0(%rsp)
vmovaps 0x1b0(%rsp), %xmm1
vmovaps 0x1a0(%rsp), %xmm2
vsubps %xmm2, %xmm1, %xmm1
vmovaps %xmm1, 0x110(%rsp)
vmovaps 0x70(%rsp), %xmm2
vmovaps 0x80(%rsp), %xmm1
vmovaps %xmm2, 0x2c0(%rsp)
vmovaps %xmm1, 0x2b0(%rsp)
vmovaps 0x2c0(%rsp), %xmm1
vmovaps 0x2b0(%rsp), %xmm2
vaddps %xmm2, %xmm1, %xmm1
vmovaps %xmm1, 0x150(%rsp)
vmovaps 0x110(%rsp), %xmm2
vmovaps 0x90(%rsp), %xmm1
vmovaps %xmm2, 0x2a0(%rsp)
vmovaps %xmm1, 0x290(%rsp)
vmovaps 0x2a0(%rsp), %xmm1
vmovaps 0x290(%rsp), %xmm2
vaddps %xmm2, %xmm1, %xmm1
vmovaps %xmm1, 0x140(%rsp)
vmovaps 0x150(%rsp), %xmm2
vmovaps 0x120(%rsp), %xmm1
vmovaps %xmm2, 0x840(%rsp)
vmovaps %xmm1, 0x830(%rsp)
vmovdqa 0x840(%rsp), %xmm1
vmovdqa 0x830(%rsp), %xmm2
vpxor %xmm2, %xmm1, %xmm1
movq 0x168(%rsp), %rax
vmovdqa %xmm1, (%rax)
vmovaps 0x140(%rsp), %xmm2
vmovaps 0xb0(%rsp), %xmm1
vmovaps %xmm2, 0x820(%rsp)
vmovaps %xmm1, 0x810(%rsp)
vmovdqa 0x820(%rsp), %xmm1
vmovdqa 0x810(%rsp), %xmm2
vpxor %xmm2, %xmm1, %xmm1
movq 0x160(%rsp), %rax
vmovdqa %xmm1, (%rax)
vmovaps -0x30(%rsp), %xmm1
vmovaps %xmm0, 0x20(%rsp)
vmovaps 0x20(%rsp), %xmm0
vmovaps %xmm1, 0x8a0(%rsp)
vmovaps %xmm0, 0x890(%rsp)
vmovaps 0x8a0(%rsp), %xmm0
vmovaps 0x890(%rsp), %xmm1
vcmpeqps %xmm1, %xmm0, %k0
vpmovm2d %k0, %xmm0
vmovaps %xmm0, -0x50(%rsp)
vmovaps -0x40(%rsp), %xmm1
vmovaps -0x50(%rsp), %xmm0
vmovaps %xmm1, 0x60(%rsp)
vmovaps %xmm0, 0x50(%rsp)
vmovaps 0x60(%rsp), %xmm0
vmovaps 0x50(%rsp), %xmm1
vpand %xmm1, %xmm0, %xmm0
vmovaps %xmm0, -0x60(%rsp)
vmovaps -0x30(%rsp), %xmm1
vmovaps -0x60(%rsp), %xmm0
vmovaps %xmm1, 0x40(%rsp)
vmovaps %xmm0, 0x30(%rsp)
vmovaps 0x40(%rsp), %xmm0
vaddps 0x30(%rsp), %xmm0, %xmm0
vmovaps %xmm0, -0x30(%rsp)
vmovaps -0x20(%rsp), %xmm1
vmovaps -0x30(%rsp), %xmm0
vmovaps %xmm1, 0x8c0(%rsp)
vmovaps %xmm0, 0x8b0(%rsp)
vmovaps 0x8c0(%rsp), %xmm0
vdivps 0x8b0(%rsp), %xmm0, %xmm0
vmovaps %xmm0, -0x70(%rsp)
vmovaps -0x70(%rsp), %xmm0
addq $0x8d8, %rsp # imm = 0x8D8
retq
nopw %cs:(%rax,%rax)
nopl (%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.