[llvm] [VPlan] Untie tail folding from optimizeInductionLiveOutUsers. NFC (PR #207984)
Luke Lau via llvm-commits
llvm-commits at lists.llvm.org
Thu Jul 9 00:24:52 PDT 2026
lukel97 wrote:
> Hm, this looks correct to me, and seems to be a good idea, although we're using more memory and cycles to achieve the same result. I'm not sure if there is some way to represent FoldTail in VPlan, but maybe Florian can comment?
There is a way to represent FoldTail in VPlan but this is purposefully trying to avoid using it, since the end value of an induction variable doesn't actually depend on whether or not the tail is folded. The aim is to avoid tying the correctness of transforms to other transforms that may or may not have run earlier in the pipeline
https://github.com/llvm/llvm-project/pull/207984
More information about the llvm-commits
mailing list