Open alex-lt-kong opened 1 year ago
I examined the below lines from cublas_utils.h:
https://github.com/NVIDIA/CUDALibrarySamples/blob/f299986ecc990c49dce295e345f9e3ad95ded188/cuBLAS/utils/cublas_utils.h#L148-L155
Given that how a matrix is generated with the same util file:
https://github.com/NVIDIA/CUDALibrarySamples/blob/f299986ecc990c49dce295e345f9e3ad95ded188/cuBLAS/utils/cublas_utils.h#L233-L238
and I am wondering shouldn't line 151 be:
std::printf("%0.2f ", A[i * lda + j]);
instead of:
std::printf("%0.2f ", A[j * lda + i]);
?
(As my understanding of cuBLAS is limited, I didn't directly raise a PR, if the issue is confirmed, I am willing to do so)
I examined the below lines from cublas_utils.h:
https://github.com/NVIDIA/CUDALibrarySamples/blob/f299986ecc990c49dce295e345f9e3ad95ded188/cuBLAS/utils/cublas_utils.h#L148-L155
Given that how a matrix is generated with the same util file:
https://github.com/NVIDIA/CUDALibrarySamples/blob/f299986ecc990c49dce295e345f9e3ad95ded188/cuBLAS/utils/cublas_utils.h#L233-L238
and I am wondering shouldn't line 151 be:
instead of:
?
(As my understanding of cuBLAS is limited, I didn't directly raise a PR, if the issue is confirmed, I am willing to do so)