[PATCH] D90554: [CostModel] remove cost-kind predicate for intrinsics in basic TTI implementation
Pengfei Wang via Phabricator via llvm-commits
llvm-commits at lists.llvm.org
Thu Nov 5 21:31:37 PST 2020
pengfei added a comment.
In D90554#2377027 <https://reviews.llvm.org/D90554#2377027>, @spatel wrote:
> Rebased after another ARM change: D90781 <https://reviews.llvm.org/D90781> / 264a6df353 <https://reviews.llvm.org/rG264a6df353b7e7ac4269bc10467bd2a991db3173>
>
> Anything else visible in here that should be fixed first?
The less unrolling in scatter-gather test seems good. But I don't have much experience in cost model. There's no objection from me.
================
Comment at: llvm/test/Analysis/CostModel/X86/intrinsic-cost-kinds.ll:46-48
; LATE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %s = call { i32, i1 } @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
; LATE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %v = call { <16 x i32>, <16 x i1> } @llvm.umul.with.overflow.v16i32(<16 x i32> %va, <16 x i32> %vb)
; LATE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret void
----------------
Why no diffs in latency?
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D90554/new/
https://reviews.llvm.org/D90554
More information about the llvm-commits
mailing list