[PATCH] D27346: X86: Split fma_pattern[_wide].ll

[Nicolai Hähnle via Phabricator via llvm-commits]

nhaehnle added a comment.

In https://reviews.llvm.org/D27346#611748, @RKSimon wrote:

> Why is this and https://reviews.llvm.org/D27347 better than just tweaking the check-prefixes?


I went off the assumption that doubling the number of RUN lines on the whole file would increase the test time more than doubling it on just a part of the file. I admit that there's the possibility that process launch overhead dominates, and I didn't actually do the timing to test it. I don't feel strongly about it in any case.


https://reviews.llvm.org/D27346