[llvm-branch-commits] [llvm] [openmp] [offload] Use HSA SVM for AMDGPU shared memory (PR #215801)

[Robert Imschweiler via llvm-branch-commits]

ro-i wrote:

Fixed the performance issue :) memset has now roughly 2x the performance with this patch compared to baseline (without this patch), so I think it's showing its benefits

https://github.com/llvm/llvm-project/pull/215801