projects / platform / upstream / llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 1698cc7) | patch
[X86] Don't disable code in combineHorizontalPredicateResult just because we have...
authorCraig Topper <craig.topper@intel.com>
Tue, 11 Feb 2020 21:14:03 +0000 (13:14 -0800)
committerCraig Topper <craig.topper@intel.com>
Tue, 11 Feb 2020 22:36:29 +0000 (14:36 -0800)
commit846d0ac43eb95c7f2f7e98b076b42c8c3208f506
treedf0e8c420d95e9fb531d718dd575e4ee6b983ab5tree | snapshot
parent1698cc7ebabe47be6cf225960e55ddb8a884d118commit | diff
[X86] Don't disable code in combineHorizontalPredicateResult just because we have avx512

We aren't doing a good job of optimizing AVX512 outside of this code. So remove the bail out for AVX512 and replace with a FIXME. This at least gets us the AVX2 codegen.

Differential Revision: https://reviews.llvm.org/D74431
llvm/lib/Target/X86/X86ISelLowering.cpp diff | blob | history
llvm/test/CodeGen/X86/vector-compare-all_of.ll diff | blob | history
llvm/test/CodeGen/X86/vector-compare-any_of.ll diff | blob | history
Domain: System / Toolchain;