Skip to content

Fix: Correct cuBLAS matrix multiplication order for Row-Major inputs #3

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
jongyunhur wants to merge 1 commit into
base: master
Choose a base branch
from
Open

Fix: Correct cuBLAS matrix multiplication order for Row-Major inputs #3

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 5 additions & 1 deletion nvml_cublas.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -104,6 +104,9 @@ void calculate( int const &m, int const &n, int const &k, nvmlClass &nvml ) {

/* Fill the matrices with test data */
/* Assume square matrices */
/* Note: We generate data in Row-Major format, but cuBLAS interprets it as Column-Major.
* This means matrices are effectively transposed when passed to cuBLAS.
* To compensate, we will compute C = B × A instead of C = A × B. */
for ( int i = 0; i < m * m; i++ ) {
h_A[i] = std::rand( ) / static_cast<data_type>( RAND_MAX );
h_B[i] = std::rand( ) / static_cast<data_type>( RAND_MAX );
Expand All @@ -124,7 +127,8 @@ void calculate( int const &m, int const &n, int const &k, nvmlClass &nvml ) {
data_type *d_C_ptr = thrust::raw_pointer_cast( &d_C[0] );

/* Performs operation using cublas */
cublasSgemm( handle, CUBLAS_OP_N, CUBLAS_OP_N, m, n, k, &alpha, d_A_ptr, lda, d_B_ptr, ldb, &beta, d_C_ptr, ldc );
/* Compute C = B * A to compensate for Row-Major to Column-Major mismatch. */
cublasSgemm( handle, CUBLAS_OP_N, CUBLAS_OP_N, n, m, k, &alpha, d_B_ptr, ldb, d_A_ptr, lda, &beta, d_C_ptr, ldc );
CUDA_RT_CALL( cudaDeviceSynchronize( ) );

/* Allocate host memory for reading back the result from device memory */
Expand Down