Fix py-scaffold unit test

[michaelmckinsey1]

Description

• Depends on py-scaffold rocm/7 fixes #1349
• Test was missing rocm=7.1.1, default rocm=6.4.3 won't work.
• Remove tioga test since we are focusing on tuolumne recently
• Need to check if gitlab test is passing on this branch after py-scaffold rocm/7 fixes #1349 is merged

[michaelmckinsey1]

The test actually runs as expected (https://lc.llnl.gov/gitlab/benchpark/benchpark/-/jobs/3819673). The issue is that after completion of the benchmark, there is always this error corrupted size vs. prev_size in fastbins, specifically on ROCm. We typically ignore this because it does not affect the benchmark run, however it fails the CI test because it then returns an exit code 134.

@joshuamvining Any ideas how to force this to exit code 0? The only thing I can think of is checking in .gitlab/utils/run-experiment.sh for exit code 134 and instead returning 0.

[joshuamvining]

The test actually runs as expected (https://lc.llnl.gov/gitlab/benchpark/benchpark/-/jobs/3819673). The issue is that after completion of the benchmark, there is always this error corrupted size vs. prev_size in fastbins, specifically on ROCm. We typically ignore this because it does not affect the benchmark run, however it fails the CI test because it then returns an exit code 134.

@joshuamvining Any ideas how to force this to exit code 0? The only thing I can think of is checking in .gitlab/utils/run-experiment.sh for exit code 134 and instead returning 0.

I think you could probably modify .gitlab/bin/exit-codes to have an exception for error code 134 whenever the benchmark is py-scaffold. That may be a cleaner place to put the exception than in run-experiment.sh