Re: [HTCondor-users] condor_submit_dag independent child processes

[Labounek Renà <xlabou01@xxxxxxxxxxxxxxxxxx>]

Small correction. I have switched processes mprage_seg and mprage2nii,  
because I have not deleted one output file from previous testing run.

The first mprage_seg is executed. After its finishing, mprage2nii  
process starts.

Other words, it is not paralelized optimally, because only 2 threads  
are executed in time when 3 independent threads can be executed.

Regards,
Rene Labounek


Cituji Labounek Renà <xlabou01@xxxxxxxxxxxxxxxxxx>:

> Dear HT-condor Users,
> I am trying to optimally paralelize our data processing pipeline via  
> condor and I have found one issue which I am not able to solve...
>
> In my file.dag, I have 6 different processes and following PARRENT  
> and CHILD edges:
>
> JOB data_convert  
> /md1/NH-DHP/NH-DHP-002/condor_logs/NH-DHP-002-data_convert.condor
> JOB dmri_preproc  
> /md1/NH-DHP/NH-DHP-002/condor_logs/NH-DHP-002-dmri_preproc.condor
> JOB mprage_bet  
> /md1/NH-DHP/NH-DHP-002/condor_logs/NH-DHP-002-mprage_bet.condor
> JOB mprage_seg  
> /md1/NH-DHP/NH-DHP-002/condor_logs/NH-DHP-002-mprage_seg.condor
> JOB mprage2diff  
> /md1/NH-DHP/NH-DHP-002/condor_logs/NH-DHP-002-mprage2diff.condor
> JOB mprage2mni  
> /md1/NH-DHP/NH-DHP-002/condor_logs/NH-DHP-002-mprage2mni.condor
> PARENT data_convert CHILD dmri_preproc mprage_bet
> PARENT mprage_bet CHILD mprage_seg mprage2mni
> PARENT dmri_preproc mprage_seg CHILD mprage2diff
>
> After data_convert, both CHILD precosses (dmri_preproc and  
> mprage_bet) start. Process dmri_preproc takes much more time than  
> other processes. Process mprage_bet is finished much earlier. When  
> mprage_bet is finnished, the process mprage2mni is started first and  
> mprage_seg is started after mprage2nii is finished and not  
> imediatelly when mprage_bet is finished.
>
> and i have enough free threads in my pool to be able to compute all  
> three processes at once.
>
> Does somebody have an idea why it behaves in that way. mprage_seg  
> and mprage2mni should be independent on themselves.
>
> Yours faithfully,
> Rene Labounek
>
> _______________________________________________
> HTCondor-users mailing list
> To unsubscribe, send a message to htcondor-users-request@xxxxxxxxxxx with a
> subject: Unsubscribe
> You can also unsubscribe by visiting
> https://lists.cs.wisc.edu/mailman/listinfo/htcondor-users
>
> The archives can be found at:
> https://lists.cs.wisc.edu/archive/htcondor-users/