Hi - I'm afraid we don't have much experience of running heavy-memory
processes with FSL under Windows - Windows in general is so much less
well-conditioned than UNIX that I'm not sure whether RAM, swap, Cygwin
tuning or other issues might be the problem here. Anyone else have more
experience in pushing the Windows limit?
Cheers, Steve.
On Mon, 25 Jul 2005, Kaufman, Galen D. wrote:
> I'm up to 2 Gb RAM now, but my PC running Windows XP with a P4 chip (3
> GHz) still won't run Melodic on a 300 volume file (128x128x40
> voxels/volume). It will do 200 volumes. The highest RAM value I've
> observed during running is ~1.4 Gb. Does anyone know any tricks to make
> Melodic run more efficiently in a Windows environment?
>
> Thanks,
>
> -----Original Message-----
> From: FSL - FMRIB's Software Library [mailto:[log in to unmask]] On
> Behalf Of Christian Beckmann
> Sent: Friday, June 17, 2005 1:26 PM
> To: [log in to unmask]
> Subject: Re: [FSL] No Melodic output
>
> Hi Galen
>
> the easiest way of finding out the required amount oof memory is to
> monitor the mem consumption of melodic using 'top' in the command line
> shell. Try running melodic on the first 100 vols and check how much mem
> is being consumed.
> cheers
> christian
>
> On 17 Jun 2005, at 14:18, Kaufman, Galen D. wrote:
>
> That did it--I 'deleted' the first 280 volumes, and got 14 components in
> the report this time.
>
> But how can I determine how much memory upgrade I need? I am currently a
> Dell PC running XP P4 chip (3 GHz), with 1 Gig RAM. What do you think
> would be a minimum set-up to crunch my 128x128x40, 300 volume files? I
> saw a file/program on the site to test the FSL installation, but can't
> find it now.
>
> Thanks a lot from a 'newbie',
> Galen
>
> -----Original Message-----
> From: FSL - FMRIB's Software Library [mailto:[log in to unmask]] On
> Behalf Of Christian Beckmann
> Sent: Friday, June 17, 2005 11:36 AM
> To: [log in to unmask]
> Subject: Re: [FSL] No Melodic output
>
> Hi Galen,
>
> It looks like you have too little memory on your system - you can test
> this by running an example melodic run on the first 20 volumes of
> filtered_func_data cheers christian
>
>
>
>
>
> On 17 Jun 2005, at 11:52, Galen Kaufman wrote:
>
> Very keen to see melodic output, but when I run the program (XP running
> Cygwin, 1 Gig RAM), the html report file has no components! I'm trying
> to evaluate a 4-D dicom format, TR=2s, 300 volumes.
>
> Any suggestions would be appreciated.
>
> The program is clearly crunching on data, then generates the following
> files:
> example_func.nii.gz (462 KB)
> filtered_func_data.nii.gz (185,942 KB)
> melodic.exe.stackdump (1 KB)
>
> The data in the big file above looks nicely extracted and aligned. But
> there is no report on any components.
>
> ...and these text files...
>
> Melodic.log:
> Melodic Version ln(11)
>
> /usr/local/fsl/bin/melodic -i filtered_func_data -o c:/Subject13_4-
> D/43031405_4_1+.ica -v --nobet --bgthreshold=10 --tr=2.0 -d 0 --
> mmthresh=0.5 --report
> ---------------------------------------------
>
> Melodic results will be in c:/Subject13_4-D/43031405_4_1+.ica
>
> Reading data file filtered_func_data ... done Create mask ... done
>
> Report.log:
> /usr/local/fsl/bin/avwroi c:/Subject13_4-D/43031405_4_1 example_func 150
> 1
>
> /usr/local/fsl/bin/slicetimer -i c:/Subject13_4-D/43031405_4_1 --
> out=prefiltered_func_data_st -r 2.0
>
> /usr/local/fsl/bin/mcflirt -in prefiltered_func_data_st -out
> prefiltered_func_data_mcf -refvol 150
>
> /usr/local/fsl/bin/betfunc prefiltered_func_data_mcf
> prefiltered_func_data_bet
>
> /usr/local/fsl/bin/ip prefiltered_func_data_bet filtered_func_data 10 -s
> 1.27388535032 -t 25.0 -1
> THRESHOLD 63.033425
>
> rm -rf prefiltered_func_data*
>
> /usr/local/fsl/bin/melodic -i filtered_func_data -o c:/Subject13_4-
> D/43031405_4_1+.ica -v --nobet --bgthreshold=10 --tr=2.0 -d 0 --
> mmthresh=0.5 --report
>
> Melodic Version ln(11)
>
> Melodic results will be in c:/Subject13_4-D/43031405_4_1+.ica
>
> Reading data file filtered_func_data ... done Create mask ... done
> Signal 6
>
>
> --
> 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
>
>
> --
> 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
>
--
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
|