[flang] [llvm] [flang][openacc] Make num_gangs, num_workers and vector_length behavior homogenous with parallel (PR #136341)

[Erich Keane via llvm-commits]

https://github.com/erichkeane approved this pull request.


https://github.com/llvm/llvm-project/pull/136341