projects / oota-llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 654a66d) | patch
InstCombe: Infer nsw for multiplies
authorDavid Majnemer <david.majnemer@gmail.com>
Fri, 26 Dec 2014 09:10:14 +0000 (09:10 +0000)
committerDavid Majnemer <david.majnemer@gmail.com>
Fri, 26 Dec 2014 09:10:14 +0000 (09:10 +0000)
commit998ae69abe6f1651ddca6046e51ff987b55e79c5
tree4753c6ae534683c1300b8fe8b5df5ab93186151btree | snapshot
parent654a66dbd362e8a570e8caea4c6a1e9e52c0c88bcommit | diff
InstCombe: Infer nsw for multiplies

We already utilize this logic for reducing overflow intrinsics, it makes
sense to reuse it for normal multiplies as well.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@224847 91177308-0d34-0410-b5e6-96231b3b80d8
lib/Transforms/InstCombine/InstCombine.h diff | blob | history
lib/Transforms/InstCombine/InstCombineAddSub.cpp diff | blob | history
lib/Transforms/InstCombine/InstCombineAndOrXor.cpp diff | blob | history
lib/Transforms/InstCombine/InstCombineMulDivRem.cpp diff | blob | history
test/Transforms/InstCombine/mul.ll diff | blob | history
C/C++ LLVM-based compilers that forbids OOTA behaviors