projects / platform / upstream / llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: cc2294a) | patch
[InstCombine] Support vector splats in foldSelectICmpAnd.
authorCraig Topper <craig.topper@intel.com>
Sat, 5 Aug 2017 20:00:41 +0000 (20:00 +0000)
committerCraig Topper <craig.topper@intel.com>
Sat, 5 Aug 2017 20:00:41 +0000 (20:00 +0000)
commit1bbcab9ca59e371f73074c052242f06cdd53038d
treeb27e089cc8d69d78c548ae0473ba1f40659d57d9tree | snapshot
parentcc2294a4ebfe628a3ced5539083494d3ecd7b51acommit | diff
[InstCombine] Support vector splats in foldSelectICmpAnd.

Unfortunately, it looks like there's some other missed optimizations in the generated code for some of these cases. I'll try to look at some of those next.

llvm-svn: 310184
llvm/lib/Transforms/InstCombine/InstCombineSelect.cpp diff | blob | history
llvm/test/Transforms/InstCombine/select-with-bitwise-ops.ll diff | blob | history
Domain: System / Toolchain;