[PATCH] D23583: [AArch64] Add feature has-fast-fma

[Evandro Menezes via llvm-commits]

evandro added inline comments.

================
Comment at: llvm/lib/Target/AArch64/AArch64ISelLowering.cpp:7378
@@ +7377,3 @@
+bool AArch64TargetLowering::enableAggressiveFMAFusion(EVT VT) const {
+  return Subtarget->hasFastFMA() && isFMAFasterThanFMulAndFAdd(VT);
+}
----------------
arsenm wrote:
> Why is this check not part of isFMAFasterThanFMulAndFAdd?
I think that they are separate things that each target should decide how to handle them.  The combiner does call `isFMAFasterThanFMulAndFAdd()` to find out if combining multiplies and sums is allowed, but it only folds some extra cases if `enableAggressiveFMAFusion()` agrees.


Repository:
  rL LLVM

https://reviews.llvm.org/D23583