File size: 4,334 Bytes
2e08f77
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
import Foundation
import Accelerate
import MetalPerformanceShaders

// Sizes of the matrices: C = A x B.
private let rowsA = 3
private let columnsA = 4
private let rowsB = columnsA
private let columnsB = 2
private let rowsC = rowsA
private let columnsC = columnsB

  private var device: MTLDevice!
  private var commandQueue: MTLCommandQueue!

  private var matrixMultiplication: MPSMatrixMultiplication!
  private var matrixA: MPSMatrix!
  private var matrixB: MPSMatrix!
  private var matrixC: MPSMatrix!

  private var arrayA = [Float](repeating: 0, count: rowsA * columnsA)
  private var arrayB = [Float](repeating: 0, count: rowsB * columnsB * 2)
  private var arrayC = [Float](repeating: 0, count: rowsC * columnsC)

  func run() {
    randomizeArrays()
    initMPS()

  }

  private func randomizeArrays() {
    // Fill up A and B with random floating point numbers (between -1 and +1).
    for i in 0..<arrayA.count {
      arrayA[i] = Float(i)
    }
    for i in 0..<arrayB.count {
      arrayB[i] = Float(i)
    }
    print(arrayB)
  }

  private func initMPS() {
    device = MTLCreateSystemDefaultDevice()
    guard device != nil else {
      fatalError("Error: This device does not support Metal")
    }

    guard MPSSupportsMTLDevice(device) else {
      fatalError("Error: This device does not support Metal Performance Shaders")
    }

    commandQueue = device.makeCommandQueue()

    matrixMultiplication = MPSMatrixMultiplication(device: device, transposeLeft: false, transposeRight: false, resultRows: rowsC, resultColumns: columnsC, interiorColumns: columnsA, alpha: 1, beta: 0)

    // For optimal speed, we should use the recommended row stride.
    //let rowBytesA = MPSMatrixDescriptor.rowBytes(fromColumns: columnsA, dataType: .float32)
    //print("preferred stride \(rowBytesA), my stride \(columnsA * MemoryLayout<Float>.stride)")

    // The contents of the arrays are copied into the MTLBuffers. Note that we
    // don't copy arrayC into bufferC because it's just zeros (arrayC is only 
    // used to store the results of the BLAS matrix multiply).
    let bufferA = device.makeBuffer(bytes: arrayA, length: rowsA * columnsA * MemoryLayout<Float>.stride, options: [])
    let bufferB = device.makeBuffer(bytes: arrayB, length: rowsB * columnsB * MemoryLayout<Float>.stride * 2, options: [])
    let bufferC = device.makeBuffer(length: rowsC * columnsC * MemoryLayout<Float>.stride, options: [])

    let descA = MPSMatrixDescriptor(dimensions: rowsA, columns: columnsA, rowBytes: columnsA * MemoryLayout<Float>.stride, dataType: .float32)
    let descB = MPSMatrixDescriptor(dimensions: rowsB, columns: columnsB, rowBytes: columnsB * MemoryLayout<Float>.stride, dataType: .float32)
    let descC = MPSMatrixDescriptor(dimensions: rowsC, columns: columnsC, rowBytes: columnsC * MemoryLayout<Float>.stride, dataType: .float32)

    matrixA = MPSMatrix(buffer: bufferA!, descriptor: descA)
    matrixB = MPSMatrix(buffer: bufferB!, offset: 0, descriptor: descB)
    matrixC = MPSMatrix(buffer: bufferC!, descriptor: descC)
      var commandBuffer = commandQueue.makeCommandBuffer()!

      matrixMultiplication.encode(commandBuffer: commandBuffer, leftMatrix: matrixA, rightMatrix: matrixB, resultMatrix: matrixC)

      commandBuffer.commit()
      commandBuffer.waitUntilCompleted()
      var contents = bufferC!.contents();
      var count = rowsA * columnsB;
    var typedPointer = contents.bindMemory(to: Float.self, capacity: count)
    var bufferedPointer = UnsafeBufferPointer(start: typedPointer, count: count)
    print(Array(bufferedPointer))

    print("Offsetted")
    matrixA = MPSMatrix(buffer: bufferA!, descriptor: descA)
    matrixB = MPSMatrix(buffer: bufferB!, offset: 4 * 2 * 4, descriptor: descB)
    matrixC = MPSMatrix(buffer: bufferC!, descriptor: descC)
      commandBuffer = commandQueue.makeCommandBuffer()!

      matrixMultiplication.encode(commandBuffer: commandBuffer, leftMatrix: matrixA, rightMatrix: matrixB, resultMatrix: matrixC)

      commandBuffer.commit()
      commandBuffer.waitUntilCompleted()
      contents = bufferC!.contents();
      count = rowsA * columnsB;
    typedPointer = contents.bindMemory(to: Float.self, capacity: count)
    bufferedPointer = UnsafeBufferPointer(start: typedPointer, count: count)
    print(Array(bufferedPointer))
  }


run()