[flang-commits] [flang] [flang] Inline minval/maxval over elemental/designate (PR #103503)

[via flang-commits]

khaki3 wrote:

Calling runtime routines might be optimal on CPUs. Perhaps, we can only apply some of the passes in OptimizedBufferization to OpenACC or OpenMP code. Any comments are welcome.


https://github.com/llvm/llvm-project/pull/103503