[PATCH] D105432: [Analysis] Add simple cost model for strict (in-order) reductions

Mon Jul 19 07:13:22 PDT 2021

david-arm added inline comments.

================
Comment at: llvm/test/Analysis/CostModel/X86/reduce-fadd.ll:146
+  %V16 = call fast float @llvm.vector.reduce.fadd.v16f32(float %arg, <16 x float> undef)
+  %V32 = call fast float @llvm.vector.reduce.fadd.v32f32(float %arg, <32 x float> undef)
   ret void
----------------
RKSimon wrote:
> please can you remove this change now that we have this test coverage below
Hi @RKSimon, the problem is that if I revert the change the tests fail. This is because we were previously calling the intrinsic without `fast`, which means that we choose the very expensive cost for an ordered reduction. If you're happy I can remove the `fast` attribute, but I will then have to update all of the costs?

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D105432/new/

https://reviews.llvm.org/D105432