Fixed Cuda Stream Protocol

[philip-paul-mueller]

In PR#980 introduced streams into the halo exchanges.
For this also DEFAULT_STREAM, which models the default stream and implements the CUDA Stream Protocol.
However, the original implementation identified as protocol version 1 instead of version 0.
Because of a related bug in GHEX this error was hidden.

This PR fixes the Python implementation and also updates GHEX.

[msimberg]

Approved, but blocking because I think we need to synchronize with a proper GHEX release given that the goal is to have stable release for icon uenvs. I'd aim to do a GHEX 0.5.1 release tomorrow.

Thank you for looking into this and fixing it!

[msimberg]

@philip-paul-mueller would you be able to test https://pypi.org/project/ghex/0.5.1/ on this PR (n.b. I managed to do the release the wrong way around, so it's now already officially on pypi, not test.pypi)?

[github-actions]

Mandatory Tests

Please make sure you run these tests via comment before you merge!

• cscs-ci run default
• cscs-ci run distributed

Optional Tests

To run benchmarks you can use:

• cscs-ci run benchmark-bencher

To run tests and benchmarks with the DaCe backend you can use:

• cscs-ci run dace

To run test levels ignored by the default test suite (mostly simple datatest for static fields computations) you can use:

• cscs-ci run extra

For more detailed information please look at CI in the EXCLAIM universe.

[philip-paul-mueller]

cscs-ci run distributed

[philip-paul-mueller]

cscs-ci run default

[philip-paul-mueller]

cscs-ci run extra

[philip-paul-mueller]

@havogt Are you fine with merging the fix?

[msimberg]

Approved, but good to wait for ok from @havogt related to branching 0.1.0 release branch.