[PATCH] D23509: [LoopVectorize] Query TTI when deciding to splat IV
Sam Parker via llvm-commits
llvm-commits at lists.llvm.org
Mon Aug 22 05:52:11 PDT 2016
samparker added a comment.
Hi Michael,
The splatting of the indvar seems to be problematic for ARM because a VDUP instruction gets generated, which has a long latency,
https://reviews.llvm.org/D23509
More information about the llvm-commits
mailing list