Hi - Mark Jenkinson is out of contact at the moment, so I'll answer this -
he can correct me later if necessary.
Yes, there seems to be a problem with flirt on Redhat8/9 which we have
been looking into - I suspect that it is a compiler bug which shows up on
certain datasets when flirt is compiled optimised.
Things should start working fine if you replace your ${FSLDIR}/bin/flirt
binary with http://www.fmrib.ox.ac.uk/fsldownloads/flirt_RH8_noopt
Make sure you make this executable after putting in place.
This is a binary compiled with no optimisation that should run (just more
slowly) on Redhat8/9.
Thanks, Steve.
On Tue, 25 Nov 2003, Hoang Tran wrote:
> Hi Mark,
>
> We ran into the FLIRT 5.0 segmentation violation issue on RedHat 9. The
> images that cause the problems are located at:
>
> http://www.nbresearch.com/data/FlirtSegViolationImages.zip
>
> This file contains two sets of images: T1 (high resolution) and ref-T1 (low
> resolution)
>
>
> The command that we used which triggers the segmentation error is:
>
> flirt -in T1 -ref ref-T1 -out foo
>
> The interesting thing is that the problem doesn't occur for every T1 image
> we have and it only occurs when trying to register the higher resolution T1
> image to the lower resolution reference image.
>
> In any case, running with verbose set at a high level shows that the
> problem always occurs just after:
>
> >> copy U UE
> >> if MAXDOF > 9
> >> clear U
> >> if MAXDOF > 9
> >> optimise 12 UE:1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 rel 2
> Coarse rotation samplings are:
> -1.570796 -0.523599 0.523599 1.570796
> -1.570796 -0.523599 0.523599 1.570796
> -1.570796 -0.523599 0.523599 1.570796 ...
>
> If any further information to debug the problem is needed, then please let
> me know.
>
> -Hoang
>
Stephen M. Smith DPhil
Associate Director, FMRIB and Analysis Research Coordinator
Oxford University Centre for Functional MRI of the Brain
John Radcliffe Hospital, Headington, Oxford OX3 9DU, UK
+44 (0) 1865 222726 (fax 222717)
[log in to unmask] http://www.fmrib.ox.ac.uk/~steve
|