[PATCH] D72524: [X86] Try to avoid casts around logical vector ops recursively.

[Florian Hahn via Phabricator via llvm-commits]

fhahn created this revision.
fhahn added reviewers: craig.topper, RKSimon, spatel.
Herald added a subscriber: hiraditya.
Herald added a project: LLVM.

Currently PromoteMaskArithemtic only looks at a single operation to
skip casts. This means we miss cases where we combine multiple masks.

This patch updates PromoteMaskArithemtic to try to recursively promote
AND/XOR/AND nodes that terminate in truncates of the right size or
constant vectors.


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D72524

Files:
  llvm/lib/Target/X86/X86ISelLowering.cpp
  llvm/test/CodeGen/X86/v8i1-masks.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D72524.237385.patch
Type: text/x-patch
Size: 63695 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20200110/850585dc/attachment-0001.bin>