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

[Matthias Braun via llvm-commits]

MatzeB added a comment.

So judging from the comments in r218120 that introduced the enableAggressiveFMAFusion() callback, this is for cases where FMAD has similar performance/resource usage as a single FADD (without even considering the FMUL). Maybe we should put that into the description of the target feature?


Repository:
  rL LLVM

https://reviews.llvm.org/D23583