Skip to content

Fix cublas caching for cublasGemv, cublasGemm

Petr Karpov requested to merge peter_fix_cublas_caching into master_pre_stage

Fix the problem with cublas caching for cublasGemv, cublasGemm.

It has been introduced with cublas (

The problem with caching was resolved by NVIDIA with cublas

For the intermediate cublas version we have to switch caching by hand using cublasLtHeuristicsCacheSetCapacity(0).

Merge request reports