[PATCH] D112760: Require 'contract' fast-math flag for FMA generation

[Matt Arsenault via Phabricator via llvm-commits]

arsenm added a comment.

I would rephrase the description as removing the global flag for contraction



================
Comment at: llvm/test/CodeGen/AMDGPU/fdot2.ll:37
 
-  %mul2 = fmul half %src1.el2, %src2.el2
-  %mul1 = fmul half %src1.el1, %src2.el1
+  %mul2 = fmul FASTMATH_FLAGS half %src1.el2, %src2.el2
+  %mul1 = fmul FASTMATH_FLAGS half %src1.el1, %src2.el1
----------------
This is somewhat surprising but I guess it works


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D112760/new/

https://reviews.llvm.org/D112760