QSIPrep on HCP-development data

Software Support
, ,
1

Hi i’m just want to make sure i’m doing this right, because eddy_cpu has been running for 2 days on 40 CPUs and it doesnt seem like it will complete anytime soon.

I’m running qsiprep 26.0.0 using singularity:

 singularity run --cleanenv \
    -B $fs_license,$bids_dir,$output_dir,$work_dir \
    $HOME/QSIprep \
    $bids_dir \
    $output_dir \
        participant \
        -w $work_dir/"$d" \
        --participant-label $d \
        --nthreads 40 \
        --omp-nthreads 40 \
        --skip-bids-validation \
        --output-resolution 2 \
        --fs-license-file $fs_license

this is the file directory structure of the single subject i’m preprocessing:

sub-0001305
├── anat
│   ├── sub-0001305_T1w.json
│   └── sub-0001305_T1w.nii.gz
└── dwi
    ├── sub-0001305_acq-dir98_dir-AP_dwi.bval
    ├── sub-0001305_acq-dir98_dir-AP_dwi.bvec
    ├── sub-0001305_acq-dir98_dir-AP_dwi.json
    ├── sub-0001305_acq-dir98_dir-AP_dwi.nii.gz
    ├── sub-0001305_acq-dir98_dir-PA_dwi.bval
    ├── sub-0001305_acq-dir98_dir-PA_dwi.bvec
    ├── sub-0001305_acq-dir98_dir-PA_dwi.json
    ├── sub-0001305_acq-dir98_dir-PA_dwi.nii.gz
    ├── sub-0001305_acq-dir99_dir-AP_dwi.bval
    ├── sub-0001305_acq-dir99_dir-AP_dwi.bvec
    ├── sub-0001305_acq-dir99_dir-AP_dwi.json
    ├── sub-0001305_acq-dir99_dir-AP_dwi.nii.gz
    ├── sub-0001305_acq-dir99_dir-PA_dwi.bval
    ├── sub-0001305_acq-dir99_dir-PA_dwi.bvec
    ├── sub-0001305_acq-dir99_dir-PA_dwi.json
    └── sub-0001305_acq-dir99_dir-PA_dwi.nii.gz
2

Hi @gerardyu,

Yes, that looks right. Eddy does take a long time and that is a lot of data per subject. If you have GPUs eddy has nice acceleration. Have you checked whether you have sufficient RAM?

Best,
Steven