Closed xjules closed 3 months ago
All modified and coverable lines are covered by tests :white_check_mark:
Project coverage is 86.01%. Comparing base (
99a29f6
) to head (6178aab
). Report is 1 commits behind head on main.
:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.
Should we also validate it in the scheduler, or is it unlikely to be called stand-alone?
NUM_CPU is not the responsibility of the Scheduler, it should be encoded into the built Realization objects, which the drivers should pick information from: https://github.com/equinor/ert/issues/8087
NUM_CPU is not the responsibility of the Scheduler, it should be encoded into the built Realization objects, which the drivers should pick information from: #8087
My comment was related to MAX_SUBMIT, which could be 0, which is the responsibility of the scheduler?
Should we also validate it in the scheduler, or is it unlikely to be called stand-alone?
Hm, do you mean like to raise ValueError
when the scheduler starts?
My comment was related to MAX_SUBMIT, which could be 0, which is the responsibility of the scheduler?
Ah! It might make sense for the scheduler to allow max_submit == 0, it serves as a "dry-run" feature (even proven to work in the wild!). But it should throw a ValueError if it is negative.
Issue Resolves https://github.com/equinor/ert/issues/7294 Resolves https://github.com/equinor/ert/issues/8078
Approach Implement positive_int validation
(Screenshot of new behavior in GUI if applicable)
When applicable