[PATCH] D116879: [llvm] Allow auto-vectorization of sincos() using libmvec

[Tim Schmielau via Phabricator via llvm-commits]

tim.schmielau added a comment.

A potential issue is that working sincos() support in libmvec <https://sourceware.org/git/?p=glibc.git;a=commit;h=ee2196bb6766ca7e63a1ba22ebb7619a3266776a> is "only" a bit over 5 years old.
E.g. a user on Centos 7 //compiling code with a vectorizable call to `sincos()` **and** compiling with `-fveclib=libmvec` **and** `-O2` or higher// will now run into an undefined reference to the vectorized `sincos()` function and will either have to deactivate auto-vectorization or update their libmvec.

Given the very specific circumstances where this applies I am not sure whether to consider that a bug or a feature - a user that much concerned with performance of numerical codes might well appreciate the change.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D116879/new/

https://reviews.llvm.org/D116879