[Openmp-commits] [openmp] [OpenMP][AMDGPU] Adapt dynamic callstack sizes to HIP behavior (PR #74080)

[Michael Halkenhäuser via Openmp-commits]

mhalk wrote:

As stated in the testcase, I would have liked to add a failing configuration, which demonstrates the impact of insufficiently sized scratch memory -- but those cases result in a hang.
If anybody has suggestions / an opinion on this, I would be very grateful.
Even when it comes to removing this run configuration altogether.

https://github.com/llvm/llvm-project/pull/74080