Hi, Dear all,
here i use the command: fmriprep-docker -w fmriprepworkdirectory/sub-01 --fs-license-file license.txt/ fmriprep1/Nifti fmriprep1/derivatives --participant_label sub-01 --skip_bids_validation --nthreads 2 --omp-nthreads 4 --mem-mb 32000
at the beginning, it was fast.
but later it stays at this point over one night, without further processing,
recon-all from Freesurfer does take a while, so you should try waiting a bit more. You do not need to explicitly set the nthreads and omp-nthreads argument. By default, fMRIPrep should use the maximum number available to maximize efficiency. You can also try to enable –resource-monitor in the fMRIPrep command to see if CPU and memory are being maxed out at all.
Hi, Steven,
Thank you for your suggestion. Ok, I just wait. But even right now, i guess over 20 hours ago, it still remains at that running point, is is really normal. Sorry for the stupid question, but i tried several times to increase the running speed, but always face such situation and have never seen how it look likes for fmriprep finishing the process. So i am curious whether anything i can do to imrove.
If the fmriprep can use the maximum number available, what the function of nthreads and n-threads are? as i understand, this is to setting to increase the running speed of processing?
Greetings
Chan
The nthreads arguments can be used to not use all of the CPUs, in case you want to conserve CPU usage. Also, and I might be mistaken, but sometimes using single threads can marginally increase reproducibility. There is a cloud based alternative called brainlife.io where you can upload your data and run apps like fMRIprep across distributed computing clusters so you aren’t reliant on your own machine. It’s a really cool site, I encourage you to check it out!
