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

[Sanjay Patel via llvm-commits]

spatel added subscribers: Gerolf, spatel.
spatel added a comment.

In https://reviews.llvm.org/D23583#518669, @rengolin wrote:

> In https://reviews.llvm.org/D23583#518632, @evandro wrote:
>
> > I disagree.  `isFMAFasterThanFMulAndFAdd()` is static and there's no good reason why `enableAggressiveFMAFusion()` cannot be static either.   In this sense, this feature is no different than the many of the existing features in AArch64.
>
>
> I assumed the FMA fusion optimisation could just query the cost model in some way. Overriding the cost model for specific instructions in specific sub-architectures is the most logical solution to me.


I was going to suggest adding to MachineCombiner, but it's already done for AArch64 with:
https://reviews.llvm.org/rL267328

There was a great list of examples detailing why FMA codegen decisions are difficult in the initial review for that commit:
https://reviews.llvm.org/D18751#402906


Repository:
  rL LLVM

https://reviews.llvm.org/D23583