Fix the behavior of extent(r) for r > rank in TeamGEMM

[yasahi-hpc]

May partially resolve #2622
After looking at the bahaviors of older versions, I found that
extent(r) for r > rank should return 1

This PR fixes the behavior.

[cwpearson]

I don't know the precedence of || and ternary offhand, could you please insert parens (even if not technically necessary)

[cwpearson]

I'd rather see this conditional nesting inverted with a constexpr check on the rank, but I won't block on it.

[lucbv]

Can we add a small unit-test that exhibit the current issue and shows that the proposed PR fixes it?

[yasahi-hpc]

I don't know the precedence of || and ternary offhand, could you please insert parens (even if not technically necessary)

Sure

[yasahi-hpc]

Can we add a small unit-test that exhibit the current issue and shows that the proposed PR fixes it?

If this is related, it means that there is a GEMM operation on rank 0 view.
I can add a test for that in #2628

[ndellingwood]

I'm building the Ifpack2 tests to check the impact of this PR, thanks for tracking this down @yasahi-hpc

[ndellingwood]

Unfortunately I still see the failures in Ifpack2_BlockTriDiContainerUnitAndPerfTests_MPI_4 when building Trilinos with this PR, tested in a Cuda build

[yasahi-hpc]

Unfortunately I still see the failures in Ifpack2_BlockTriDiContainerUnitAndPerfTests_MPI_4 when building Trilinos with this PR, tested in a Cuda build

Thanks a lot for testing and reporting.
That is a bit unfortunate.
I will continue investigation

[ndellingwood]

Hi @yasahi-hpc , I retested with 9f1b00a in a Cuda build but I am still seeing failures with the Ifpack2_BlockTriDiContainerUnitAndPerfTests_MPI_4 test

[yasahi-hpc]

Hi @yasahi-hpc , I retested with 9f1b00a in a Cuda build but I am still seeing failures with the Ifpack2_BlockTriDiContainerUnitAndPerfTests_MPI_4 test

Thank you for testing.
That is unfortunate.

I will investigate on my side.
First of all, I need to build Trilinos on my environment.

[yasahi-hpc]

@lucbv Can I close this and start working again #2628 and #2651 ?

[lucbv]

Yes, I think we can, sorry we took a while to make the final decision on how to move forward.
The code changes in GEMM led to issues with CUDA and HIP so we might want to re-introduce things more slowly, with more smaller PRs...

[yasahi-hpc]

Thank you for the information
So, better not to work on #2628 and #2651