[PATCH] D38417: [test-suite] Adding HACCKernels app

[Amara Emerson via Phabricator via llvm-commits]

aemerson added a comment.

In https://reviews.llvm.org/D38417#887050, @homerdin wrote:

> I made the change that Hal mentioned to allow for the loop to vectorize.


Could you add a comment to the source to explain why there's a difference in pragmas, for future reference? Ideally we want to minimize as many clang specific vectorization pragmas, if the vectorizer isn't working for whatever reason then it should be noticed, especially if there's a regression somewhere in piping AA information through.


https://reviews.llvm.org/D38417