Resource overuse Issues with (clinica) T1-Linear Pipeline

[Shahzad ALi]

Hi Clinica Community,
I’m running the t1-linear and t1-volume pipeline on my Mac M3 Pro (18GB RAM, 12-core CPU) and facing memory/CPU overuse issues, causing the pipeline to halt.
Could anyone suggest the possible solution?

Thanks in advance!
Best,

Shahzad 

----------------------

[Shahzad ALi]

pipeline.log:

  Cannot allocate job 6843 (0.20GB, 1 threads).
250307-11:51:26,658 nipype.workflow DEBUG:
Cannot allocate job 6844 (0.20GB, 1 threads).
250307-11:51:26,658 nipype.workflow DEBUG:
Cannot allocate job 6845 (0.20GB, 1 threads).
250307-11:51:26,658 nipype.workflow DEBUG:
Cannot allocate job 6846 (0.20GB, 1 threads).
250307-11:51:26,658 nipype.workflow DEBUG:
Cannot allocate job 6847 (0.20GB, 1 threads).
250307-11:51:26,658 nipype.workflow DEBUG:
Cannot allocate job 6848 (0.20GB, 1 threads).
250307-11:51:26,658 nipype.workflow DEBUG:
Cannot allocate job 6849 (0.20GB, 1 threads).
250307-11:51:26,658 nipype.workflow DEBUG:
Cannot allocate job 6850 (0.20GB, 1 threads).
250307-11:51:26,658 nipype.workflow DEBUG:
Cannot allocate job 6851 (0.20GB, 1 threads).
250307-11:51:26,658 nipype.workflow DEBUG:
Cannot allocate job 6852 (0.20GB, 1 threads).

Best,

-----------------------------------------------------------------------

[Alice Joubert]

Hello,

In order for us to help you as best as possible, could you detail your issue a little bit more ?

- Is this happening for both T1 Linear and T1 Volume separately ?

- Are you running processes simultaneously ?

- Are you using any specific options for these pipelines that might affect the CPU usage ? For example : in T1-Linear --use-antspy or in any of both the option --n-procs ? (see the documentation for explanations about options)

- Do you know at which point of the pipeline these errors happen ? (Example : during the processing or at the writing step)

Best,

Alice

--
You received this message because you are subscribed to the Google Groups "Clinica" group.
To unsubscribe from this group and stop receiving emails from it, send an email to clinica-user...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/clinica-user/CAPaAoQTm77aMY90MHSwWELChgn6N1ASB9YqaVN7-qMTKzw-7sQ%40mail.gmail.com.

[Shahzad ALi]

Hi, Dear Alice,

Thanks for you reply, Here are replies to your questions:

• - Is this happening for both T1 Linear and T1 Volume separately ?

• currently, I tried for t1-linear only.

• - Are you running processes simultaneously ?

• Separately

• - Are you using any specific options for these pipelines that might affect the CPU usage ? For example : in T1-Linear --use-antspy or in any of both the option --n-procs ? (see the documentation for explanations about options)

• i tried using these command (with and without --n_procs)
• clinica run t1-linear BIDS_Converted/T1 CAPS_DIR --random_seed 12
• clinica run t1-linear BIDS_Converted/T1 CAPS_DIR --random_seed 12 —n_procs 3

• - Do you know at which point of the pipeline these errors happen ? (Example : during the processing or at the writing step)

• during writing step, after processing

Best,

Shahzad Ali 

-----------------------------------------------------------------------

To view this discussion visit https://groups.google.com/d/msgid/clinica-user/CAP_epFvM8shw4QrdWg7K_tLPcM%3DLYQPh%2BHcsek2JxWD4QO7hKw%40mail.gmail.com.

[Alice Joubert]

Hello,

Sorry for the late reply. If you still have this issue, possible solutions are :

- reducing the number of cores allocated to the job

- using a working directory to keep intermediary results and relaunch easily the job if it fails

Hope that helps,

Alice

To view this discussion visit https://groups.google.com/d/msgid/clinica-user/CAPaAoQRCwtwaWFHj%3DSRd8ONm8dCtnkcJWKSJ1%2BAw2hsUg1xNLA%40mail.gmail.com.