summaryrefslogtreecommitdiffstats
path: root/llvm/test/CodeGen/AMDGPU/v1024.ll
Commit message (Collapse)AuthorAgeFilesLines
* [AMDGPU] Change register type for v32 vectorsStanislav Mekhanoshin2019-07-161-0/+29
When it is AReg_1024 this results in unnecessary copying into AGPRs of a 32 element vectors even though they are not intended for an mfma instruction. Differential Revision: https://reviews.llvm.org/D64815 llvm-svn: 366252
OpenPOWER on IntegriCloud