Given that that looks like HCP data, it might be that you are running into
a bug with melodicıs memory allocation of large files.
Peace,
Matt.
On 3/1/17, 9:19 PM, "FSL - FMRIB's Software Library on behalf of Jenny
Gilbert" <[log in to unmask] on behalf of [log in to unmask]> wrote:
>Hi FSL gurus,
>
>I am trying to complete ICA, but keep running into an error where melodic
>runs out of memory soon after launching the command. Currently, I am
>testing ICA with 10 subjects, but plan to scale up to 334, which is why I
>included the flags for MIGP in the command. However, melodic keeps
>crashing with the 10 subject test. Can anyone help me figure out how to
>fix this? I've included my output including the error message below.
>
>Thank you in advance,
>Jenny
>
>----------------------------------------------------- Output
>-----------------------------------------------------
>Melodic results will be in /projects/niblab/data/twin_ICA_test.ica
>
>Create mask ... done
>Randomising input file order
>Reading data file /projects/niblab/data/subjects/1230/rest1 ... done
> Estimating data smoothness ... done
>
>terminate called after throwing an instance of 'std::bad_alloc'
> what(): std::bad_alloc
>/projects/niblab/scripts/twin/ICA_test.sh: line 1: 27531 Aborted
> melodic -i /projects/niblab/scripts/twin/ICA_test_input.txt --migp
>--sep_vn -o /projects/niblab/data/twin_ICA_test.ica -v --report --nobet
>--bgthreshold=10 --tr=0.720 -d 50 --mmthresh=0.5 --Oall
|