Use Kokkos::abort directly

[masterleinad]

Part of #14751. This reverts some changes in #14753.
CUDA and HIP device functions can only be used in a different translation unit if special flags are passed that allow this. However, allowing relocatable device code oftentimes prevents optimization. In particular, it doesn't seem worth it if we would only need it for kokkos_abort. Thus, this pull request uses Kokkos::abort directly at the cost of including Kokkos_Core.hpp. An alternative would be to just ignore Assert in kernels executed on the device.

[masterleinad]

/rebuild

[masterleinad]

/rebuild

[masterleinad]

/rebuild

[tjhei]

/rebuild

It looks like the MPI build might be hanging. Can this be related to this PR?

[masterleinad]

It looks like the MPI build might be hanging. Can this be related to this PR?

This pull request should only inline the call to Kokkos::abort so it shouldn't change any existing behavior.

[bangerth]

This is of course a bit disappointing, because Kokkos_Core.hpp pulls in many thousands of lines of code. Is the expected increase in compile time measurable?

[masterleinad]

This is of course a bit disappointing, because Kokkos_Core.hpp pulls in many thousands of lines of code. Is the expected increase in compile time measurable?

We are trying to address that in the next Kokkos release by providing a separate header file.

Locally I am seeing,

real	9m23.509s
user	40m42.230s
sys	2m1.631s

for a debug build of the library and

real	9m7.645s
user	39m56.241s
sys	1m55.335s

without this patch.

[bangerth]

That seems acceptable to me. I'm ok with merging, but maybe we can get a second vote.

[bangerth]

Thank you also for testing this out!

[kronbichler]

I think the increase in compile time is not nice, but acceptable, so let us move forward here.