Bennet and Steve, thanks for your advice. For the moment, i am not using slspec and I should have a alot of RAM (32) and I not running any othr processes. And as some data that initially gets a segviol later passes through, the error seems so stochastic.
Looking at my bvec-file, I have like 10 decimals for each direction, how many do you use? could this be related?
Niklas Lenfeldt, PhD, MSc E.P
Umeå University
Phone: +46706096687
________________________________________
Från: FSL - FMRIB's Software Library <[log in to unmask]> för Bennet Fauber <[log in to unmask]>
Skickat: den 20 november 2018 19:55
Till: [log in to unmask]
Ämne: Re: [FSL] SV: [FSL] segmentation error
We recently had problems with segfaulting when there was a mismatch
between the number of lines in the slspec file and the number of
volumes in the image file.
Is your data on a network file server? Any possibility that there is
some transient problem reading the data from any of the files? Also
possible as Steve Smith observed faulty RAM or exceeding RAM+swap.
On Sat, Nov 17, 2018 at 3:39 AM Niklas Lenfeldt <[log in to unmask]> wrote:
>
> Thanks. it is very strange. Some of the data that yielded a segemenation error first...now went through the entire loop without any problem
>
> and the error does not only occur the first time, it can be another iteration.
>
> so I do not think there is an error in the b-files. I have checked with dtifit that L1 fits with the fibers. Not for all data but the dicom to nifti conversion is the same for all data.
>
> when is the new eddy due?
>
> best nick
>
> Niklas Lenfeldt, PhD, MSc E.P
> Umeå University
> Phone: +46706096687
>
> ________________________________________
> Från: FSL - FMRIB's Software Library <[log in to unmask]> för Jesper Andersson <[log in to unmask]>
> Skickat: den 9 november 2018 15:06
> Till: [log in to unmask]
> Ämne: Re: [FSL] segmentation error
>
> Dear Niklas,
>
> it is hard to say exactly what has gone wrong from the messages below. It seems it goes wrong when estimating the parameters for the Gaussian Process for the first time. One possible reason for that can be if there is a mismatch between the information in the bvals/bvecs file and what is actually encoded in the images, i.e. if one has gotten the bvals/bvecs files wrong.
>
> If that is not the problem I would need to take a look at your data myself. But there is no point in me doing that before the 6.0.1 release of eddy, to first see if maybe that fixes your problem.
>
> Jesper
>
>
> > On 31 Oct 2018, at 09:47, Niklas Lenfeldt <[log in to unmask]> wrote:
> >
> > while running eddy on a few of my samples, I get a segmentation violation
> >
> > [niklasl@localhost test]$ ./preproc_eddy3T.txt
> > Reading images
> > Performing volume-to-volume registration
> > Running Register
> > Loading prediction maker
> > Evaluating prediction maker model
> > Segmentation violation, Address not mapped, Offending address = 0x770000006e
> > /lib64/libc.so.6 [0x7ff023f2b3c1]
> > eddy ) [0x6062eb] [
> > eddy ) [0x5ff163] [
> > eddy ) [0x45fab5] [K�
> > ./preproc_eddy3T.txt: line 4: 21758 Segmentation fault (core dumped) eddy --imain="${katalog}_DTI.nii" --mask="${katalog}_DTI_brain_mask.nii.gz" --acqp="${katalog}_DTI_acqparam.txt" --index="${katalog}_DTI_index.txt" --bvecs="${katalog}_DTI.bvec" --bvals="${katalog}_DTI.bval" --out="${katalog}_DTI_corr" --slm=linear --repol --niter=5 --fwhm=10,0,0,0,0 -v
> >
> >
> > does anyone have an idea? It happens only on a few cases. the rest works fine. have opened them in fsleys and I think they look no worse than other samples.
> >
> > best
> >
> > nick
> >
> >
> > ########################################################################
> >
> > To unsubscribe from the FSL list, click the following link:
> > https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1
>
>
> ########################################################################
>
> To unsubscribe from the FSL list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1
>
>
> ########################################################################
>
> To unsubscribe from the FSL list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1
########################################################################
To unsubscribe from the FSL list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1
########################################################################
To unsubscribe from the FSL list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1
|