Update slurm tests returned perf_values

[jgphpc]

• follow-up of Return perf_value as float #375
• Run with:

UENV=prgenv-gnu/25.06:rc5 ./R -c cscs-reframe-tests.git/checks/system/slurm/slurm.py -J p=normal -r

• You can manually trigger 1 (or more) CI pipelines by adding a cscs-ci run alps-daint-uenv comment in the Pull Request
• By default all UENVs will be tested, however you can restrict to a single UENV with: cscs-ci run alps-daint-uenv;MY_UENV=cp2k/2025.1:v2

[gppezzi]

@jgphpc the tests are passing on Daint, but are you addressing the gres issue reported on SD-66684?

It is unclear, as the old test (which you are deleting in this PR) still fails and the updated one (SlurmGPUGresTest) doesn't.

[jgphpc]

@jgphpc the tests are passing on Daint, but are you addressing the gres issue reported on SD-66684?

It is unclear, as the old test (which you are deleting in this PR) still fails and the updated one (DefaultRequestGPUSetsGRES) doesn't.

• I did not delete gres_gpu.py, I moved it into slurm.py (no reason to keep it separate).
• This PR is not really addressing the issue in SD-66684, it's more about sending the right data to Elastic (I updated the tests. in the file while I was reading it).

[jgphpc]

https://reframetalk.slack.com/archives/CF8ENLUDB/p1756210315486089

[gppezzi]

@jgphpc the tests are passing on Daint, but are you addressing the gres issue reported on SD-66684?
It is unclear, as the old test (which you are deleting in this PR) still fails and the updated one (DefaultRequestGPUSetsGRES) doesn't.

* I did not delete gres_gpu.py, I moved it into slurm.py (no reason to keep it separate).

* This PR is not really addressing the issue in SD-66684, it's more about sending the right data to Elastic (I updated the tests. in the file while I was reading it).

ok now I found the SlurmGPUGresTest (I was looking at the wrong gres test), but my question remains: all tests in this PR pass now on daint, while the old gres_gpu.py fails.

Which is odd if this PR is not addressing SD-66684, but it could be me running it wrong?

[gppezzi]

maybe leave a comment to document why you are changing this?