[PATCH] D156112: [AArch64][LoopVectorize] Improve tail-folding heuristic on neoverse-v1

[Paul Walker via Phabricator via llvm-commits]

paulwalker-arm added a comment.

If for this function we cannot produce a better heuristic based souly on VF and the number of instructions then we're really talking about needing to extend the cost model itself and ensure that such a model correctly costs all the predicated logic, which includes the while based control flow.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D156112/new/

https://reviews.llvm.org/D156112