]> granicus.if.org Git - llvm/commit
projects / llvm / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 9aae96d) | patch
[X86] Consider isCodeGenOnly in the EVEX2VEX pass to make VMAXPD/PS map to the non...
authorCraig Topper <craig.topper@intel.com>
Tue, 1 Oct 2019 07:10:09 +0000 (07:10 +0000)
committerCraig Topper <craig.topper@intel.com>
Tue, 1 Oct 2019 07:10:09 +0000 (07:10 +0000)
commitf7e6d9fc311578b22265fd0ebe3a1f6a90fcbc75
tree7ad000d63eb02b73f2034f88f5c0cc1a317b1129tree | snapshot
parent9aae96d0508550bdf4d45f6c28cf1613ab5c1ca0commit | diff
[X86] Consider isCodeGenOnly in the EVEX2VEX pass to make VMAXPD/PS map to the non-commutable VEX instruction. Use EVEX2VEX override to fix the scalar instructions.

Previously the match was ambiguous and VMAXPS/PD and VMAXCPS/PD
were mapped to the same VEX instruction. But we should keep
the commutableness when change the opcode.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@373303 91177308-0d34-0410-b5e6-96231b3b80d8
lib/Target/X86/X86InstrAVX512.td diff | blob | history
test/CodeGen/X86/evex-to-vex-compress.mir diff | blob | history
utils/TableGen/X86EVEX2VEXTablesEmitter.cpp diff | blob | history
Unnamed repository; edit this file 'description' to name the repository.