[PATCH] D70951: [Matrix] Use fmuladd when lowering matrix.multiply.

[Adam Nemet via Phabricator via llvm-commits]

anemet added inline comments.


================
Comment at: llvm/lib/Transforms/Scalar/LowerMatrixIntrinsics.cpp:50
+             "result in different results, due to less rounding error."));
+
 namespace {
----------------
fhahn wrote:
> anemet wrote:
> > Shouldn't this be using the fpflags from the intrinsic?  I.e. I was thinking of a clang flag.
> We can certainly add this on the clang level, although I am not sure how easy/hard it will be to add additional flags there.
Can we do both?  I.e. propagate fp-contract from the intrinsic's fpflag and also allow the backend to globally enable this?  I'd be fine with that.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D70951/new/

https://reviews.llvm.org/D70951