[llvm] LV]: consider scalable VF during deciding dead epilogue. (PR #156724)
David Sherwood via llvm-commits
llvm-commits at lists.llvm.org
Mon Nov 10 06:56:39 PST 2025
================
@@ -4407,6 +4407,13 @@ VectorizationFactor LoopVectorizationPlanner::selectEpilogueVectorizationFactor(
LLVM_DEBUG(dbgs() << "LEV: Maximum Trip Count for Epilogue: "
<< MaxTripCount << "\n");
}
+ // Check if the RemainingIterations is scalable.
----------------
david-arm wrote:
It's about how common it is, but more about the complexity of the code and the missing cases. Im my opinion, I think if we proceed with this approach there still needs to be some changes because I don't think it really makes sense to try comparing values in different spaces without using estimates.
https://github.com/llvm/llvm-project/pull/156724
More information about the llvm-commits
mailing list