[Bug libgomp/122280] target teams distribute parallel for collapse(2) yields different results in a matmul than separate loops (one with omp target teams distribute the second with omp parallel for) on nvptx target. Clang compiles the code correctly

[burnus at gcc dot gnu.org via Gcc-bugs]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=122280

--- Comment #12 from Tobias Burnus <burnus at gcc dot gnu.org> ---
(In reply to Benjamin Schulz from comment #9)
> Created attachment 62693 [details]
> main.cpp
> 
> I now reduced the test case to a single file, which does 3 matrix
> multiplications and distplays the result. 

I have now run it 20 times with identical output and successful. I really
wonder what does wrong for you. I probably should later test this on a
different system and see whether it reproduces there.