 |
 |
Dear Matthew,
Thanks a lot for your answer. The randomise_parallel never worked in my case, so I am just running 20 processes one per each core, I just open 20 tmux and run in each one of them one randomise command. In this way I parallelise my work on the CPUs.
After the reinstallation I could not replicate my results anymore, but the files I sent you are the same I have use before the reinstallation and also the data. I have checked again the design matrix and the contrasts and everything seems fine to me.
I would like to ask you if it can be possible that before even if the FSL version was the same v6.0, some packages were from the old version of FSL and randomise while running was using dependencies from old version of FSL.
The thing is that now the randomise function runs faster than before and other packages like eddy have more functions than before the reinstallation. Now I have eddy_cuda8.0 and eddy_cuda9.1 before even if the version of FSL was 6.0, these function were not there.
I am in the meanwhile working on a project with DTI data and using eddy_openmp, even for this data after the reinstallation of FSL the results of the eddy_openmp changed. For this reason I was quite confused by the situation because in both projects I saw different results with the same data, same scripts after reinstalling FSL.
I am really thankful for your time and your help in understanding what could have been the issue.
Best,
Susanna
University of Zurich
Department of Economics
Blümlisalpstrasse 10, Room BLU-104
8006 Zurich
Phone: +41 44 634 55 61
Email: [log in to unmask]
From: Matthew Webster
Sent: 06 June 2019 15:27
To: [log in to unmask]
Subject: Re: [FSL] FSL randomise results changed after re-installing FSL
Dear Susanna,
I’ve been unable to replicate your “before” results ( my test replicates the “after” results ). How did you parallelise randomise on your cloud machine ( e.g. with randomise_parallel or some other method )? The results in ROIMilkModelNoFIbeforeFSLreinstall are quite strange: the *tstat1 images match the “after” results, but the _tstat7 “before” image also matches _tstat1 which implies the same contrast is being processe in both cases. This would imply some kind of structured issue in the before results.
Hope this helps,
Kind Regards
Matthew
--------------------------------
Dr Matthew Webster
FMRIB Centre
John Radcliffe Hospital
University of Oxford
> On 5 Jun 2019, at 15:31, Susanna Gobbi <[log in to unmask]> wrote:
>
> Dear Matthew,
> I have just upload the data.
>
> Thanks a lot for the help,
> Best,
>
> Susanna
>
> University of Zurich
> Department of Economics
> Blümlisalpstrasse 10, Room BLU-104
> 8006 Zurich
>
> Phone: +41 44 634 55 61
> Email: [log in to unmask]
>
>
> From: Matthew Webster
> Sent: 05 June 2019 14:47
> To: [log in to unmask]
> Subject: Re: [FSL] FSL randomise results changed after re-installing FSL
>
> Dear Susanna,
> Can you upload your original data (MilkshakeAll.nii) as well ( so we can replicate the analysis ) to:
>
> https://oxfile.ox.ac.uk/oxfile/work/extBox?id=700839B35190E48BF
>
> Kind Regards
> Matthew
> --------------------------------
> Dr Matthew Webster
> FMRIB Centre
> John Radcliffe Hospital
> University of Oxford
>
> > On 5 Jun 2019, at 10:32, Susanna Gobbi <[log in to unmask]> wrote:
> >
> > Dear Taylor,
> > Thanks a lot for your time and help. I have upload a folder called files with an example of the same analysis before and after the FSL reinstallation.
> >
> > Here my command
> > randomise -i ./SusannaStudy/MilkshakeAll.nii -o ./SusannaStudy/ResultsModel3/Milkshake/ROI/ -d ./SusannaStudy/designNoFI.mat -t ./SusannaStudy/designNoFI.con -m rROImask.nii -e ./SusannaStudy/design.grp -x --uncorrp -v 6 -T
> > Best,
> >
> > Susanna
> >
> > University of Zurich
> > Department of Economics
> > Blümlisalpstrasse 10, Room BLU-104
> > 8006 Zurich
> >
> > Phone: +41 44 634 55 61
> > Email: [log in to unmask]
> >
> >
> > From: Taylor Hanayik
> > Sent: 05 June 2019 10:26
> > To: [log in to unmask]
> > Subject: Re: [FSL] FSL randomise results changed after re-installing FSL
> >
> > Hi Susanna,
> >
> > We are looking into this some more. Would you be able to upload your “before” and “After” results for your randomise analysis? (After meaning after you updated FSL).
> >
> > You can upload to Oxford’s file sharing service here:
> >
> > https://oxfile.ox.ac.uk/oxfile/work/extBox?id=738200123033B05CFC
> >
> > Could you also reply back with the command you are using to start the randomise program?
> >
> > Thanks!
> >
> >
> > Cheers,
> >
> > Taylor Hanayik
> >
> >
> > Taylor Hanayik PhD
> > Analysis Research Software Engineer
> > FMRIB, John Radcliffe Hospital
> > University of Oxford
> >
> >
> >
> >
> >
> >
> > On 4 Jun 2019, at 13:28, Susanna Gobbi <[log in to unmask]> wrote:
> >
> > https://oxfile.ox.ac.uk/oxfile/work/extBox?id=738200123033B05CFC
> >
> >
> > <8ED21B4141914192B2A6E9B4D639F53C.png>
> > 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
> >
> > <8ED21B4141914192B2A6E9B4D639F53C.png>
>
>
> ########################################################################
>
> 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