[PATCH] D96805: [AMDGPU][CostModel] Refine cost model for control-flow instructions.

[Stanislav Mekhanoshin via Phabricator via llvm-commits]

rampitec added a comment.

AMDGPU estimates look reasonable, but bumping unroll threshold twice looks suspicious. Is that to accommodate some loops with switches?
I am afraid this can have quite unpredictable impact on overall performance. Did you run any perf tests?


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D96805/new/

https://reviews.llvm.org/D96805