You say its interpolated up to 240x240x130? So the true resolution is
coarser than your pixel size? Could you not resample your data back to the
original resolution, then run bedpost? This would certainly be faster and
less expensive than buying a new computer. Or is there anything unsound
about down-sampling data? I would imagine you'd end up losing a bit of
resolution.
Drew
On Fri, 24 Nov 2006 12:09:26 +0000, Martin Kavec <[log in to unmask]> wrote:
>Thanks for reply Tim.
>
>It's 240x240x130 and 15 directions plus b0. I forgot to keen an eye on the
>interpolation so the reconstructed in-plane voxel size is unnecessarily
>small. Thanks,
>
>Martin
>
>
>On Friday 24 November 2006 09:34, Tim Behrens wrote:
>> What are your image dimensions? It sounds like you are running out of
>> RAM in the merge.
>> Even if we were able to solve this problem, I suspect you may have
>> similar problems when it comes to tractography.
>>
>> On 24 Nov 2006, at 08:56, Martin Kavec wrote:
>> > Hi,
>> >
>> > I have quite large DTI data (240 MB), which causes bedpost to crash
>> > (fails to allocate memory) while merging results (thsamples,
>> > phsamples) at the end of analysis. Exact error message is and is
>> > repeated for all the failed files:
>> >
>> > avwmerge(29235) malloc:*** vm_allocate(size=1497600000)
>> > failed (error code=3)
>> > avwmerge(29235) malloc:***
>> > error: can't allocate region
>> > avwmerge(29235) malloc:***
>> > set a breakpoint in szone_error to debug
>> > /opt/fsl/bin/bedpost: line 256:
>> > 29253 Bus error avwmerge -z ${subj}.bedpost/merged_thsamples
>> > `${FSLDIR}/bin/imglob -oneperimage ${subj}.bedpost/dif_slices/
>> > data_slice_*/thsamples`
>> >
>> > I run fsl-3.3 on a MacBook Pro with 2GB of memory. I am not sure if
>> > I am hitting a roof of the system's capabilities to allocate the
>> > memory or I simply don't have enough of one. Nevertheless, I would
>> > be really happy if I could have the results. Is there any way to
>> > merge the results "manually"?
>> >
>> > Thanks a lot for help,
>> >
>> > Martin
>
>--
>*********************************
>Senior Research Associate in Neuroradiology
>MR Unit
>Erasme Hospital
>808 Route de Lennik
>B-1070 Brussels
>BELGIUM
>*********************************
>=========================================================================
|