diff options
| author | Simon Pilgrim <llvm-dev@redking.me.uk> | 2021-05-27 15:36:29 +0100 |
|---|---|---|
| committer | Simon Pilgrim <llvm-dev@redking.me.uk> | 2021-05-27 16:07:42 +0100 |
| commit | fe8d97cbe5ecec50731650947b4e3b45f49228f8 (patch) | |
| tree | 21e8a61628f3800120a8b374904309c0226521c0 /clang/lib/Frontend/CompilerInvocation.cpp | |
| parent | e49d6e16235ac48e4dc55535a571989925b8da56 (diff) | |
| download | llvm-fe8d97cbe5ecec50731650947b4e3b45f49228f8.zip llvm-fe8d97cbe5ecec50731650947b4e3b45f49228f8.tar.gz llvm-fe8d97cbe5ecec50731650947b4e3b45f49228f8.tar.bz2 | |
[CostModel][X86] AVX512 truncation ops are slower than cost models indicate.
The SkylakeServer model (and later IceLake/TigerLake targets according to Agner) have the PMOV truncations as uops=2, rthroughput=2 instructions.
Noticed while trying to reduce the diffs between cost tables and llvm-mca analysis.
Diffstat (limited to 'clang/lib/Frontend/CompilerInvocation.cpp')
0 files changed, 0 insertions, 0 deletions