#261 Improve MC computation to have all procs available running jobs.

[gautierbureau]

Checklist before requesting a review

use '[x]' to check the checkboxes, or submit the PR and then click the checkboxes

• unit tests and non regression tests were added (update of algorithms)
• main documentation was updated (update of input/output file, update of algorithms)
• if this PR modifies a dictionary: the corresponding french dictionary was updated
• if this commit targets a release branch: the corresponding milestone was added in the ticket and in this PR

[rosiereflo]

Please add a nrt that exercise this

[gautierbureau]

Please add a nrt that exercise this

By design all nrts already exercise this. And no way to check it from an nrt point of view as this behavior is not always active if at some point there is not enough jobs to run...

[rosiereflo]

Please add a nrt that exercise this

By design all nrts already exercise this. And no way to check it from an nrt point of view as this behavior is not always active if at some point there is not enough jobs to run...

the log of the MC should reflect that

[gautierbureau]

Please add a nrt that exercise this

By design all nrts already exercise this. And no way to check it from an nrt point of view as this behavior is not always active if at some point there is not enough jobs to run...

the log of the MC should reflect that

There are no logs as reference

[rosiereflo]

Please add a nrt that exercise this

By design all nrts already exercise this. And no way to check it from an nrt point of view as this behavior is not always active if at some point there is not enough jobs to run...

the log of the MC should reflect that

There are no logs as reference

No, but it might be the right time to add it

[gautierbureau]

Please add a nrt that exercise this

By design all nrts already exercise this. And no way to check it from an nrt point of view as this behavior is not always active if at some point there is not enough jobs to run...

the log of the MC should reflect that

There are no logs as reference

No, but it might be the right time to add it

I am not sure we have something to compare those two logs...

[rosiereflo]

Please add a nrt that exercise this

By design all nrts already exercise this. And no way to check it from an nrt point of view as this behavior is not always active if at some point there is not enough jobs to run...

the log of the MC should reflect that

There are no logs as reference

No, but it might be the right time to add it

I am not sure we have something to compare those two logs...

The log comparison already exists in dynawo, the only things to do is to add the logs in reference and modify the filter in the LineToCompare function in nrtDiff.py

[gautierbureau]

Please add a nrt that exercise this

By design all nrts already exercise this. And no way to check it from an nrt point of view as this behavior is not always active if at some point there is not enough jobs to run...

the log of the MC should reflect that

There are no logs as reference

No, but it might be the right time to add it

I am not sure we have something to compare those two logs...

The log comparison already exists in dynawo, the only things to do is to add the logs in reference and modify the filter in the LineToCompare function in nrtDiff.py

I will add the logs file prior to this PR to see a difference. #308

[rosiereflo]

It seems that the nrt are passing even without updating the logs which is surprising

[gautierbureau]

It seems that the nrt are passing even without updating the logs which is surprising

I think we don't use enough procs to activate the behavior...

[rosiereflo]

It seems that the nrt are passing even without updating the logs which is surprising

I think we don't use enough procs to activate the behavior...

I think we use 2, we need a test of this behavior somehow to avoid silent regressions

[gautierbureau]

I agree