Could cublas do pinned memory allocation?

0
votes

I understand that pinned memory allocated by "cudaHostAlloc" can be transferred more efficiently to device than "malloc"'ed memory can. However, I think "cudaHostAlloc" can only be compiled by cuda compiler. My scenario is to use cublas API without cuda compiler, and it seems that cublas doesn't provide function for pinned memory allocation from the handbook, or maybe I miss something...

1
cudacublas
I am not sure this is an appropriate topic to be posted at stackoverflow, and if not, please close this question and I will ask on nVidia forum. - Hailiang Zhang
Just realized that I only did "#include <cublas_v2.h>", but cudaHostAlloc is defined in <cublas.h>. It is a little bit confusing that a later version (I assume for "v2") doesn't contain the previous features. - Hailiang Zhang

1 Answers

3
votes

cudaHostAlloc() is implemented in the CUDA Runtime API. You don't need to compile with nvcc to use CUDA API calls, you can just include the appropriate header (e.g. cuda_runtime_api.h) and link with the runtime library (cudart).