0
我想写一个使用cuBLAS函数cublasSgemm的矩阵乘法的简单例子。我的代码如下所示:在这段代码中使用cublasSgemm不成功的矩阵乘法的原因是什么?
int m =100, n = 100;
float * bold1 = new float [m*n];
float * bold2 = new float [m*n];
float * bold3 = new float [m*n];
for (int i = 0; i< m; i++)
for(int j = 0; j <n;j++)
{
bold1[i*n+j]=rand()%10;
bold2[i*n+j]=rand()%10;
}
cudaError_t cudaStat;
cublasStatus_t stat;
cublasHandle_t handle;
const float alpha = 1.0;
const float beta = 0;
float * dev_bold1, * dev_bold2, *dev_bold3;
cudaStat = cudaMalloc ((void**)&bold1, sizeof(float)*m*n);
if(cudaStat != CUBLAS_STATUS_SUCCESS)
{
cout<<"problem1";
return cudaStat;
}
cudaStat = cudaMalloc ((void**)&bold2,sizeof(float)*m*n);
if(cudaStat != CUBLAS_STATUS_SUCCESS)
{
cout<<"problem2";
return cudaStat;
}
cudaStat = cudaMalloc ((void**)&bold3,sizeof(float)*m*n);
if(cudaStat != CUBLAS_STATUS_SUCCESS)
{
cout<<"problem3";
return cudaStat;
}
cublasSetMatrix(m,n,sizeof(float),bold1,m,dev_bold1,m);
cublasSetMatrix(m,n,sizeof(float),bold2,m,dev_bold2,m);
stat = cublasCreate(&handle);
if(stat != CUBLAS_STATUS_SUCCESS)
{
cout<<"problem4";
return stat;
}
cout<<stat<<" "<<CUBLAS_STATUS_SUCCESS<<"\n";
stat = cublasSgemm(handle, CUBLAS_OP_N, CUBLAS_OP_N, m, m, n ,&alpha, dev_bold1, n, dev_bold2, n, &beta,dev_bold3,m);
if (stat != CUBLAS_STATUS_SUCCESS)
{
cout<<"problem5";
return stat;
}
cudaStat = cudaMemcpy(bold3,dev_bold3,sizeof(float)*m*n,cudaMemcpyDeviceToHost);
if (cudaStat != cudaSuccess)
{
cout<<"problem6";
return cudaStat;
}
delete []bold1;
delete []bold2;
cudaFree(dev_bold1);
cudaFree(dev_bold2);
cudaFree(dev_bold3);
在此代码中,我想乘以矩阵bold1和bold2填充随机数。 代码返回“问题5”这是关系到这部分代码:
stat = cublasSgemm(handle, CUBLAS_OP_N, CUBLAS_OP_N, m, m, n ,&alpha, dev_bold1, n, dev_bold2, n, &beta,dev_bold3,m);
if (stat != CUBLAS_STATUS_SUCCESS)
{
cout<<"problem5";
return stat;
}
我还打印统计并显示“13”!
任何人都可以请帮我理解我的代码有什么问题吗? 谢谢!