Hello Bernd,
As run_first_all submits its sub-jobs to the queue, you would need to modify $FSLDIR/bin/run_first_all to use the appropriate job IDs/halts to ensure your fslstats call is run at the right time.
Kind Regards
Matthew
> Hi to all and happy new year,
>
> I have an additional question using fsl_sub and run_first_all:
> I want to calculate hippocampal volumes on 50 patients.
> Therefore I have written a bash script which looks like this:
>
> for j in patientDir
> do
>
> cd $j
> /usr/local/fsl/bin/run_first_all -s L_Hipp,R_Hipp -i mprage.nii -o fsl_seg_output.nii
>
> done
>
> But now I would like to complete this step (in the script) immediately by using
>
> fslstats output -L_Hipp_corr.nii -V
>
> which is not possible (in my opinion), as I do not know, when the run_first_all command has finished. And otherwise fslstats would result in an error, as the script starts fslstats without a regular input.
> Therefore my question would be how could I run this for-loop on a multi-core-cluster (>50), so that each patient data can be processed separately ? Maybe I missed something how to use fsl_sub.
>
> Many thanks,
> Regards, Bernd
|