[PATCH] D32596: [DAGCombine] Transform (fmul X, -2.0) --> (fneg (fadd X, X)).

[Chad Rosier via Phabricator via llvm-commits]

mcrosier updated this revision to Diff 97131.
mcrosier retitled this revision from "[InstCombine] Transform fmul X, +2.0 --> fadd X, X." to "[DAGCombine] Transform (fmul X, -2.0) --> (fneg (fadd X, X)).".
mcrosier edited the summary of this revision.
mcrosier added a comment.
Herald added a subscriber: nhaehnle.

-Rewrite as a DAG combine.


https://reviews.llvm.org/D32596

Files:
  lib/CodeGen/SelectionDAG/DAGCombiner.cpp
  test/CodeGen/AArch64/fmul-combines.ll
  test/CodeGen/AMDGPU/fmul-2-combine-multi-use.ll
  test/CodeGen/AMDGPU/fmuladd.f32.ll
  test/CodeGen/X86/fmul-combines.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D32596.97131.patch
Type: text/x-patch
Size: 4628 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20170428/affd67a9/attachment.bin>