[flang-commits] [flang] [flang][mlir][fir] Unroll inner loops in presence of user vectorizati… (PR #210820)

[Steve Scalpone via flang-commits]

sscalpone wrote:

Nice results!  I just skimmed the changes.
1. Is this change using `mlir::loopUnrollFull`?  If not, why not?
2. Is this a workaround until the llvm vectorizer can handle the inner loop?  If so, is there a test to tell us when llvm is capable so we can revisit this change?

https://github.com/llvm/llvm-project/pull/210820