there are no hard-coded limitations - it all depends on the hardware.
Have you tried monitoring the process with 'top' to see how much mem it
tries to allocate?
On 18 Feb 2004, at 20:24, Stan Colcombe wrote:
> Hi All-
> I have some fMRI data sent to me from a collaborator, in a
> 128x128x20x912 (x,y,z,t) matrix (total file size ~500megs). mcflirt
> and ip worked fine, but when running film_gls the program aborted
> without any explicit error message, and only the following: 'Loading
> image...Aborted.' I pulled out a smaller ROI (64 64 20 912), which ran
> fine. So my *guess* is that there is a problem in running such a large
> image through film_gls, no? (I'm running a dual 2Ghz Linux system,
> with 2G RAM and 2G swap, but got the same failure on a dual G4 Mac,
> also with 2G RAM).
> Before proceeding, I was hoping that someone could tell me a) if the
> problem is with image size & film_gls, and if so, b) where that limit
> was (total image size, xy matrix size, etc). I apologize in advance if
> I have missed this in the docs somewhere. One thought is to warp the
> series to MNI, then re-run the analysis; but while that decreases the
> xy matrix dims it increases the total file size. Or, I could split the
> file before film, etc. Other thoughts?
Christian F. Beckmann
Oxford University Centre for Functional
Magnetic Resonance Imaging of the Brain,
John Radcliffe Hospital, Headington, Oxford OX3 9DU, UK
Email: [log in to unmask] - http://www.fmrib.ox.ac.uk/~beckmann/
Phone: +44(0)1865 222782 Fax: +44(0)1865 222717