[flang-commits] [flang] [flang][HLFIR] Optimize MINLOC/MAXLOC for equality masks (PR #186916)

[via flang-commits]

anoopkg6 wrote:

Thank you for the suggestion. Moving this to InlineElementals after expansion provides a generalized approach to inline the hlfir.apply of the mask into the reduction loop. This can eliminate temporary array creation for all masked intrinsics, rather than being limited to specific equality comparisons.

https://github.com/llvm/llvm-project/pull/186916