[Mlir-commits] [clang] [clang-tools-extra] [libcxx] [lldb] [llvm] [mlir] [CodeGen] Don't scalarize scalable math intrinsics that have a veclib mapping (PR #216309)

[Mattéo Rizza Murgier]

matteo-rm wrote:

@arsenm Thank you for taking a look at this. Closing this PR: the branch history got tangled, and the approach is more of a workaround. An actual fix using RTLIB seems to require continuing the work started in #167026 which is a much bigger piece of work than I thought.

https://github.com/llvm/llvm-project/pull/216309