[PATCH] D31169: [DAGCombiner] Initial support for the fast-math flag contract

[Adam Nemet via Phabricator via llvm-commits]

anemet created this revision.

Now alternatively to the TargetOption.AllowFPOpFusion global flag, FMUL->FADD
can also use the per operation FMF to allow fusion.

The idea here is not to port everything to the new scheme (e.g. fused
multiply-and-sub will be ported later) but that this work all the way from
clang.

The transformation is conditionalized on *both* the FADD and the FMUL having
the FMF contract flag.


https://reviews.llvm.org/D31169

Files:
  lib/CodeGen/SelectionDAG/DAGCombiner.cpp
  test/CodeGen/AArch64/neon-fma-FMF.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D31169.92425.patch
Type: text/x-patch
Size: 7109 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20170321/4a63e2e4/attachment.bin>