Hi again Björn,
sorry for the wait. The reason is that I have been unable to repeat your problem. I have run both the CPU and the GPU versions on the data you sent me, without any crashes.
Can you please run one of these data sets with the --very_verbose flag so that we can have a little more info on what might go wrong?
Jesper
On 22 Apr 2016, at 08:26, Björn Åström <[log in to unmask]> wrote:
> Hi Jesper
>
> Yes, from what I've seen it's the same error every time. I have also
> tried to convert again from DICOM to nifti but no change.
>
> The command is pretty much the basic eddy, I pasted the full script
> I'm using in my first post.
>
> eddy --imain=NYPUM_206_008_30T_V01_DTI.nii
> --mask=NYPUM_206_008_30T_V01_DTI_bet_015_mask.nii
> --acqp=NYPUM_206_008_30T_V01_acq.txt
> --index=NYPUM_206_008_30T_V01_index.txt
> --bvecs=NYPUM_206_008_30T_V01_DTI.bvec
> --bvals=NYPUM_206_008_30T_V01_DTI.bval
> --out=NYPUM_206_008_30T_V01_eddy --slm=linear
>
> 4 example files that give out errors have been uploaded through the
> link you provided, named "bjorn_astrom_160421.tar.gz"
>
> Thanks again!
>
> Björn Åström
> Department of clinical neuroscience
> Umea University
>
> On 4/21/16, Jesper Andersson <[log in to unmask]> wrote:
>> Dear Björn,
>>
>> is the error consistent? I.e., if you were to run your script multiple times
>> would the same subjects cause problems each time? If so, can you please
>> upload an example where you have this problem? Please make a tar-ball
>> containing all files alluded to be the eddy command line and upload it to
>> https://oxfile.ox.ac.uk/oxfile/work/extBox?id=72139C7E463068D5F
>>
>> Jesper
>>
>> On 19 Apr 2016, at 09:54, Björn Åström
>> <[log in to unmask]<mailto:[log in to unmask]>> wrote:
>>
>> Hi FSL experts
>>
>> I'm currently running eddy and getting this error message (see below) on
>> some of my subjects (roughly 1/10, meaning that 9/10 are processed
>> properly), all files are formatted and named in exactly the same fashion
>> among all subjects. I've seen that other people have had similar error
>> messages in previous FSL-posts, possibly related to bval and/or bvec files?
>> I have not noticed any discernible discrepancies between the bvecs of those
>> analysis that work and those who don't.
>>
>> $ ./eddy_prog_final160417.txt
>> 0 1000 1000 1000 1000 1000 1000 1000 1000 1000 1000 1000 1000 1000 1000
>> 1000
>> 16
>> 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
>> 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1Segmentation violation, Unknown reason,
>> Offending address = (nil)
>> eddy ) [0x5e0826] [
>> eddy ) [0x5db62b] [
>> eddy ) [0x43401b] [
>> ./eddy_prog_final160417.txt: line 2: 5846 Segmentation fault (core
>> dumped) eddy --imain="${katalog}_DTI.nii"
>> --mask="${katalog}_DTI_bet_015_mask.nii" --acqp="${katalog}_acq.txt"
>> --index="${katalog}_index.txt" --bvecs="${katalog}_DTI.bvec"
>> --bvals="${katalog}_DTI.bval" --out="${katalog}_eddy" --slm=linear
>>
>> This is my script:
>>
>> #!/bin/bash
>> for kat_slash in */ #
>> /home/lenf/test_fslpreproc/glm_160415/NYPUM_083_005_30T_V01
>> do
>> katalog=$(basename $kat_slash)
>> cd $katalog
>> fslroi "${katalog}_DTI.nii" "${katalog}_DTI_B0" 0 1 # B0-volymen
>> bet "${katalog}_DTI_B0" "${katalog}_DTI_bet_015" -m -f 0.15 # bet pa
>> B0-volymen
>> printf "1 0 0 0.05" > "${katalog}_acq.txt" # skriv acq-filen
>> grad_vol=($(cat ${katalog}_DTI.bval)) # las in bvec i array grad_vol
>> echo ${grad_vol[*]}
>> echo ${#grad_vol[*]}
>> var_index="${grad_vol[@]//*0/1}" # ersatt varje b-value i grad_vol med 1
>>> lika manga 1:or som volymer
>> printf "${var_index}" > ${katalog}_index.txt # skriv ut var_index med
>> ettor till indexfilen
>> echo $var_index
>> cat "${katalog}_index.txt"
>> eddy --imain="${katalog}_DTI.nii"
>> --mask="${katalog}_DTI_bet_015_mask.nii" --acqp="${katalog}_acq.txt"
>> --index="${katalog}_index.txt" --bvecs="${katalog}_DTI.bvec"
>> --bvals="${katalog}_DTI.bval" --out="${katalog}_eddy" --slm=linear
>> cd ..
>> done
>>
>> Any help on this would be greatly appreciated
>>
>> Kind regards
>>
>> Björn Åström
>> Department of clinical neuroscience
>> Umea University
>>
>>
|