maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity

[X86] AMD Zen 3 has fast variable per-lane shuffles 

... but lane-crossing shuffles are slow.
This commit is contained in:
Roman Lebedev 2021-06-01 10:46:05 +03:00
parent cf9b1f7a0e
commit a3b8695bf5
No known key found for this signature in database
GPG Key ID: 083C3EBB4A1689E0
1 changed files with 3 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
llvm/lib/Target/X86/X86.td
View File

@ -1111,7 +1111,9 @@ def ProcessorFeatures {
FeaturePKU,
FeatureVAES,
FeatureVPCLMULQDQ];
list<SubtargetFeature> ZN3AdditionalTuning = [FeatureMacroFusion];
list<SubtargetFeature> ZN3AdditionalTuning =
[FeatureMacroFusion,
FeatureFastVariablePerLaneShuffle];
list<SubtargetFeature> ZN3Tuning =
!listconcat(ZNTuning, ZN3AdditionalTuning);
list<SubtargetFeature> ZN3Features =