[PATCH] D33983: update add\sub costs of vectors of 64 in X86\SLM arch

[Mohammed Agabaria via Phabricator via llvm-commits]

magabari added a comment.

fixed test comment



================
Comment at: test/Transforms/LoopVectorize/X86/slm-no-vectorize.ll:1
+; This test should not be vectorized in X86\SLM arch
+; this test has a bad performance (regression of -70%) if vectorized on SLM arch
----------------
efriedma wrote:
> The vectorizer is clearly doing something wrong if it's generating a 64-bit multiply for this loop... might want to clarify the comment.
fixed


https://reviews.llvm.org/D33983