[PATCH] D12096: [InstCombineAddSub opportunities]: More opportunities to factorize FAdd/FSub when unsafeAlgebra is present for Inst

[Aditya Nandakumar via llvm-commits]

aditya_nandakumar updated this revision to Diff 32784.
aditya_nandakumar added a comment.

Updated to use hasOneUse
Consistent comments.

The IR I see in the test case is a simplified version of running reassociate followed by instcombine - the original IR contained lots of -ve floating point constants which reassociate now canonicalizes further expecting instcombine to cleanup. Instcombine is unable to simplify possibly due how reassociate arranges the expressions (my guess).


http://reviews.llvm.org/D12096

Files:
  lib/Transforms/InstCombine/InstCombineAddSub.cpp
  test/Transforms/InstCombine/FAddFSubAssociativeFactorize.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D12096.32784.patch
Type: text/x-patch
Size: 17652 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20150821/aed36233/attachment.bin>