Function MatMulKernel

code/Chapter12/12-6-mpi-hip-cannon.cpp:9–20

#define DIMS 2000
#define BLOCK_SIZE 32
__global__ void MatMulKernel(float *A, float *B, float *C, int m1, int n1, int n2)
{
    int y = hipBlockIdx_y * hipBlockDim_y + hipThreadIdx_y;
    int x = hipBlockIdx_x * hipBlockDim_x + hipThreadIdx_x;
    if(x < n2 && y < m1){
        int tmp = 0;
        for(int n = 0; n < n1; n++){
            tmp += A[y * n1 + n] * B[n * n2 + x];
        }
        C[y * n2 + x] += tmp;
    }
}
void  Mul_Matrixdcu(float *A, float *B, float *C, int m1, int n1, int n2)
{
    float *d_A, *d_B, *d_C;