[flang-commits] [flang] [mlir] [mlir][flang][openmp] Rework parallel reduction operations (PR #79308)

[Kiran Chandramohan via flang-commits]

kiranchandramohan wrote:

> Could a new test be added that would show the new capabilities?

I have added a new test test/Lower/OpenMP/parallel-reduction.f90 (https://github.com/llvm/llvm-project/pull/79308/files#diff-a659fd2e2c1970fcce5baa536110c82fbc58e30c522b8e36c769cd6eb8383436). The test should print the number of threads. Previously this test case would not work correctly since it always expected the reduction operation to be present in the region. Now we have removed this restriction allowing more cases to be handled and to match the spec better.

If there are no concerns, I would like to rebase and submit this.

https://github.com/llvm/llvm-project/pull/79308