TiledMatrixMul.cu

/**
 * Copyright 1993-2015 NVIDIA Corporation.  All rights reserved.
 *
 * Please refer to the NVIDIA end user license agreement (EULA) associated
 * with this source code for terms and conditions that govern your use of
 * this software. Any use, reproduction, disclosure, or distribution of
 * this software and related documentation outside the terms of the EULA
 * is strictly prohibited.
 *
 */

/**
 * Matrix multiplication: C = A * B.
 * Host code.
 *
 * This sample implements matrix multiplication which makes use of shared memory
 * to ensure data reuse, the matrix multiplication is done using tiling approach.
 * It has been written for clarity of exposition to illustrate various CUDA programming
 * principles, not with the goal of providing the most performant generic kernel for matrix multiplication.
 * See also:
 * V. Volkov and J. Demmel, "Benchmarking GPUs to tune dense linear algebra,"
 * in Proc. 2008 ACM/IEEE Conf. on Supercomputing (SC '08),
 * Piscataway, NJ: IEEE Press, 2008, pp. Art. 31:1-11.
 */

// System includes
#include <stdio.h>
#include <assert.h>
#include<string> //z

// CUDA runtime
#include <cuda_runtime.h>

// Helper functions and utilities to work with CUDA
#include <helper_functions.h>
#include <helper_cuda.h>

#include<stdlib.h>


/**
 * Matrix multiplication (CUDA Kernel) on the device: C = A * B
 * wA is A's width and wB is B's width
 */
template <int BLOCK_SIZE> __global__ void MatrixMulCUDA(float *C, float *A, float *B, int wA, int wB) 
{
    if(wA!=wB) //z
    {
        printf("A and B are not both square matrix\n"); //z
        assert(0); //z
    }
    // Block index
    int bx = blockIdx.x;
    int by = blockIdx.y;

    // Thread index
    int tx = threadIdx.x;
    int ty = threadIdx.y;

    // Index of the first sub-matrix of A processed by the block
    int aBegin = wA * BLOCK_SIZE * by;

    // Index of the last sub-matrix of A processed by the block
    //z int aEnd   = aBegin + wA - 1;

    // Step size used to iterate through the sub-matrices of A
    int aStep  = BLOCK_SIZE;

    // Index of the first sub-matrix of B processed by the block
    int bBegin = BLOCK_SIZE * bx;

    // Step size used to iterate through the sub-matrices of B
    int bStep  = BLOCK_SIZE * wB;

    // Csub is used to store the element of the block sub-matrix
    // that is computed by the thread
    float Csub = 0;

    int row = wA; //z

    int XY_BLOCKS = (BLOCK_SIZE+row-1) /BLOCK_SIZE; //z
    // Loop over all the sub-matrices of A and B
    // required to compute the block sub-matrix
    //z for (int a = aBegin, b = bBegin;
    //z        a <= aEnd;
    //z        a += aStep, b += bStep) {
    int a = aBegin, b = bBegin; //z
        // Declaration of the shared memory array As used to
        // store the sub-matrix of A
        __shared__ float As[BLOCK_SIZE][BLOCK_SIZE];

        // Declaration of the shared memory array Bs used to
        // store the sub-matrix of B
        __shared__ float Bs[BLOCK_SIZE][BLOCK_SIZE];

        // Load the matrices from device memory
        // to shared memory; each thread loads
        // one element of each matrix
        
        //z As[ty][tx] = A[a + wA * ty + tx];
        //z Bs[ty][tx] = B[b + wB * ty + tx];

        for(int i=0;i<XY_BLOCKS;i++) //z
        {
            if((i*BLOCK_SIZE+tx)<row &&(by*BLOCK_SIZE+ty)<row) //z
            {
                As[ty][tx] = A[a + row * ty + tx]; //z
            }
            else{
                As[ty][tx] = 0; //z
            }
            if((i*BLOCK_SIZE+ty)<row &&(bx*BLOCK_SIZE+tx)<row) //z
            {
                Bs[ty][tx] = B[b + row * ty + tx];//z
            }
            else{
                Bs[ty][tx] = 0; //z
            }
            a += aStep; //z
            b += bStep; //z

        // Synchronize to make sure the matrices are loaded
        __syncthreads();

        // Multiply the two matrices together;
        // each thread computes one element
        // of the block sub-matrix
#pragma unroll

        for (int k = 0; k < BLOCK_SIZE; ++k) {
            Csub += As[ty][k] * Bs[k][tx];
        }

        // Synchronize to make sure that the preceding
        // computation is done before loading two new
        // sub-matrices of A and B in the next iteration
        __syncthreads();
        }

        //z // Synchronize to make sure the matrices are loaded
        //z __syncthreads();

        // Multiply the two matrices together;
        // each thread computes one element
        // of the block sub-matrix
//z #pragma unroll

    //z    for (int k = 0; k < BLOCK_SIZE; ++k) {
       //z     Csub += As[ty][k] * Bs[k][tx];
        //z }

        // Synchronize to make sure that the preceding
        // computation is done before loading two new
        // sub-matrices of A and B in the next iteration
       //z  __syncthreads();
    //z }

    // Write the block sub-matrix to device memory;
    // each thread writes one element
     //z int c = wB * BLOCK_SIZE * by + BLOCK_SIZE * bx;
     //z C[c + wB * ty + tx] = Csub; //TODO
     if((BLOCK_SIZE*bx+tx)<row && ((BLOCK_SIZE*by+ty)<row))
    {
        C[row * (BLOCK_SIZE * by+ty) + (BLOCK_SIZE * bx + tx) ] = Csub;
    }
}

void ConstantInit(float *data, int size, float val) {
    for (int i = 0; i < size; ++i) {
        data[i] = val;
    }
}

/**
 * Run a simple test of matrix multiplication using CUDA
 */
int MatrixMultiply(int argc, char **argv,
                   int block_size, const dim3 &dimsA,
                   const dim3 &dimsB) {
    // Allocate host memory for matrices A and B
    unsigned int size_A = dimsA.x * dimsA.y;
    unsigned int mem_size_A = sizeof(float) * size_A;
    float *h_A = reinterpret_cast<float *>(malloc(mem_size_A));
    unsigned int size_B = dimsB.x * dimsB.y;
    unsigned int mem_size_B = sizeof(float) * size_B;
    float *h_B = reinterpret_cast<float *>(malloc(mem_size_B));

    // Initialize host memory
    const float valB = 0.01f;
    ConstantInit(h_A, size_A, 1.0f);
    ConstantInit(h_B, size_B, valB);

    // Allocate device memory
    float *d_A, *d_B, *d_C;

    // Allocate host matrix C
    dim3 dimsC(dimsB.x, dimsA.y, 1);
    unsigned int mem_size_C = dimsC.x * dimsC.y * sizeof(float);
    float *h_C = reinterpret_cast<float *>(malloc(mem_size_C));

    if (h_C == NULL) {
        fprintf(stderr, "Failed to allocate host matrix C!\n");
        exit(EXIT_FAILURE);
    }

    checkCudaErrors(cudaMalloc(reinterpret_cast<void **>(&d_A), mem_size_A));

    checkCudaErrors(cudaMalloc(reinterpret_cast<void **>(&d_B), mem_size_B));

    checkCudaErrors(cudaMalloc(reinterpret_cast<void **>(&d_C), mem_size_C));

    // copy host memory to device
    checkCudaErrors(cudaMemcpy(d_A, h_A, mem_size_A, cudaMemcpyHostToDevice));

    checkCudaErrors(cudaMemcpy(d_B, h_B, mem_size_B, cudaMemcpyHostToDevice));

    // Setup execution parameters
    dim3 threads(block_size, block_size);
    //z dim3 grid(dimsB.x / threads.x, dimsA.y / threads.y);

    dim3 grid((dimsB.x+block_size-1)/threads.x , (dimsA.y+block_size-1)/threads.y);

    // Create and start timer
    printf("Computing result using CUDA Kernel...\n");

    // Performs warmup operation using TiledMatrixMul CUDA kernel
    if (block_size == 16) {
        MatrixMulCUDA<16> <<< grid, threads >>>(d_C, d_A, d_B,
                                                dimsA.x, dimsB.x);
    } else {
        MatrixMulCUDA<32> <<< grid, threads >>>(d_C, d_A, d_B,
                                                dimsA.x, dimsB.x);
    }

    printf("done\n");

    cudaDeviceSynchronize();

    // Allocate CUDA events that we'll use for timing
    cudaEvent_t start;
    checkCudaErrors(cudaEventCreate(&start));

    cudaEvent_t stop;
    checkCudaErrors(cudaEventCreate(&stop));

    // Record the start event
    checkCudaErrors(cudaEventRecord(start, NULL));

    // Execute the kernel
    int nIter = 300;

    for (int j = 0; j < nIter; j++) {
        if (block_size == 16) {
            MatrixMulCUDA<16> <<< grid, threads >>>(d_C, d_A, d_B,
                                                    dimsA.x, dimsB.x);
        } else {
            MatrixMulCUDA<32> <<< grid, threads >>>(d_C, d_A, d_B,
                                                    dimsA.x, dimsB.x);
        }
    }

    // Record the stop event
    checkCudaErrors(cudaEventRecord(stop, NULL));

    // Wait for the stop event to complete
    checkCudaErrors(cudaEventSynchronize(stop));

    float msecTotal = 0.0f;
    checkCudaErrors(cudaEventElapsedTime(&msecTotal, start, stop));

    // Compute and print the performance
    float msecPerMatrixMul = msecTotal / nIter;
    double flopsPerMatrixMul = 2.0 * static_cast<double>(dimsA.x) *
                               static_cast<double>(dimsA.y) *
                               static_cast<double>(dimsB.x);
    double gigaFlops = (flopsPerMatrixMul * 1.0e-9f) /
                       (msecPerMatrixMul / 1000.0f);
    printf(
        "Performance= %.2f GFlop/s, Time= %.3f msec, Size= %.0f Ops," \
        " WorkgroupSize= %u threads/block\n",
        gigaFlops,
        msecPerMatrixMul,
        flopsPerMatrixMul,
        threads.x * threads.y);

    // Copy result from device to host
    checkCudaErrors(cudaMemcpy(h_C, d_C, mem_size_C, cudaMemcpyDeviceToHost));

    printf("Checking computed result for correctness: ");
    bool correct = true;

    // test relative error by the formula
    //     |<x, y>_cpu - <x,y>_gpu|/<|x|, |y|>  < eps
    double eps = 1.e-6;  // machine zero

    for (int i = 0; i < static_cast<int>(dimsC.x * dimsC.y); i++) {
        double abs_err = fabs(h_C[i] - (dimsA.x * valB));
        double dot_length = dimsA.x;
        double abs_val = fabs(h_C[i]);
        double rel_err = abs_err / abs_val / dot_length;

        if (rel_err > eps) {
            printf("Error! Matrix[%05d]=%.8f, ref=%.8f error term is > %E\n",
                   i, h_C[i], dimsA.x * valB, eps);
            correct = false;
        }
    }

    printf("%s\n", correct ? "Result = PASS" : "Result = FAIL");

    // Clean up memory
    free(h_A);
    free(h_B);
    free(h_C);
    checkCudaErrors(cudaFree(d_A));
    checkCudaErrors(cudaFree(d_B));
    checkCudaErrors(cudaFree(d_C));

    printf("\nNOTE: The CUDA Samples are not meant for performance"\
           "measurements. Results may vary when GPU Boost is enabled.\n");

    if (correct) {
        return EXIT_SUCCESS;
    } else {
        return EXIT_FAILURE;
    }
}


/**
 * Program main
 */
int main(int argc, char **argv) {
    
    std::string Argv1(argv[1]); //z
    std::string Argv2(argv[2]); //z
    int row = atoi(argv[2]); //z

    if(argc!=3 || strcmp(argv[1],"-i")!=0 || row <=0){ //z
        printf("- Usage: ./TiledMatrixMul -i <rowDim>\n"); //z
        if(row<=0) printf("- Usage: <rowDim> can't less than 1\n"); //z
        return -1; //z
    } //z


    

    

    //z printf("[Matrix Multiply Using CUDA] - Starting...\n");

    //zif (checkCmdLineFlag(argc, (const char **)argv, "help") ||
       //z     checkCmdLineFlag(argc, (const char **)argv, "?")) {
        //z printf("Usage -device=n (n >= 0 for deviceID)\n");
        //z printf("      -wA=WidthA -hA=HeightA (Width x Height of Matrix A)\n");
        //z printf("      -wB=WidthB -hB=HeightB (Width x Height of Matrix B)\n");
        //z printf("  Note: Outer matrix dimensions of A & B matrices" \
               " must be equal.\n");

        //z exit(EXIT_SUCCESS);
    //z}

    // This will pick the best possible CUDA capable device, otherwise
    // override the device ID based on input provided at the command line
    //z int dev = findCudaDevice(argc, (const char **)argv);

    int block_size = 32;

    //z dim3 dimsA(5 * 2 * block_size, 5 * 2 * block_size, 1);
    //z dim3 dimsB(5 * 4 * block_size, 5 * 2 * block_size, 1);

    dim3 dimsA(row, row, 1); //z
    dim3 dimsB(row, row, 1); //z
    // width of Matrix A
    //z if (checkCmdLineFlag(argc, (const char **)argv, "wA")) {
    //z    dimsA.x = getCmdLineArgumentInt(argc, (const char **)argv, "wA");
    //z }

    // height of Matrix A
    //z if (checkCmdLineFlag(argc, (const char **)argv, "hA")) {
    //z    dimsA.y = getCmdLineArgumentInt(argc, (const char **)argv, "hA");
    //z }

    // width of Matrix B
    //z if (checkCmdLineFlag(argc, (const char **)argv, "wB")) {
    //z    dimsB.x = getCmdLineArgumentInt(argc, (const char **)argv, "wB");
    //z }

    // height of Matrix B
    //z if (checkCmdLineFlag(argc, (const char **)argv, "hB")) {
    //z     dimsB.y = getCmdLineArgumentInt(argc, (const char **)argv, "hB");
    //z }

    // if (dimsA.x != dimsB.y) {
    //z     printf("Error: outer matrix dimensions must be equal. (%d != %d)\n",
    //z            dimsA.x, dimsB.y);
    //z     exit(EXIT_FAILURE);
    //z }

    printf("MatrixA(%d,%d), MatrixB(%d,%d)\n", dimsA.x, dimsA.y,
                                               dimsB.x, dimsB.y);

    int matrix_result = MatrixMultiply(argc, argv, block_size, dimsA, dimsB);

    exit(matrix_result);
}