 |
 |
Hi MarkI m sure not why would this be a limit for my system. I m using 4GB 1.3 GHz system. If we assume that this is the reason 7GB calculations cant be done, but then removing just 3-4 subjects make it work which is still more than the 4GB ram. And I think RAM hasnt really much to do with this issue here, a lesser RAM might result in slow processing but it still should be able to compute. Anyway to avoid this confusion, I think I will just try running it on a 16 GB machine !!Best,Qasim Qasim BukhariResearch Assistant and Doctoral CandidateInstitute for Biomedical EngineeringETH and University ZurichWolfgang-Pauli-Strasse 27, HIT E22webpages: MR-pagehttp://www.biomed.ee.ethz.ch/people/sybukhar Date: Mon, 30 Jun 2014 05:18:20 +0000 From: [log in to unmask] Subject: Re: [FSL] melodic segmentation fault To: [log in to unmask] Hi, You can calculate the memory required by multiplying the number of voxels by the number of timepoints and the number of subjects and 8 (as we use double-precision numbers which take 8 bytes each). When I do this calculation I get almost exactly 7GB for this dataset, so is it possible that this amount of memory is the limit for your system? All the best, Mark On 29 Jun 2014, at 23:55, Qasim Bukhari <[log in to unmask]> wrote: Hi Mark The dimensions are 60 x 90 x 10; time series is 350; so in total 60 x 90 x 10 x 350. I just took some images out to make it work, and now its fine. Just that I couldnt consider my whole dataset :) Anyway, thanks a lot Best, Qasim Qasim Bukhari Research Assistant and Doctoral Candidate Institute for Biomedical Engineering ETH and University Zurich Wolfgang-Pauli-Strasse 27, HIT E22 webpages: MR-page http://www.biomed.ee.ethz.ch/people/sybukhar Date: Fri, 27 Jun 2014 07:33:21 +0000 From: [log in to unmask] Subject: Re: [FSL] melodic segmentation fault To: [log in to unmask] Hi Qasim, Out of interest, how large were your images? It might be possible to calculate why you have this limitation. All the best, Mark On 24 Jun 2014, at 22:57, Qasim Bukhari <[log in to unmask]> wrote: Hi everybody, I found a solution to the problem. Atleast in my case. Its very strange what I found. Basically in concat ICA it doesnt take more than 47 subjects at a time. The problem is I had 50 subjects to input and as soon as I remove any 3 subjects; "ANY three" from the list of files, then it works. So now then; is there any solution to actually consider all 50 subjects; may be a way around or something ?? Otherwise I ll just discard 3 subjects if there is no other way. Best, Qasim Qasim Bukhari Research Assistant and Doctoral Candidate Institute for Biomedical Engineering ETH and University Zurich Wolfgang-Pauli-Strasse 27, HIT E22 webpages: MR-page http://www.biomed.ee.ethz.ch/people/sybukhar > Date: Mon, 23 Jun 2014 15:40:19 +0200 > From: [log in to unmask] > Subject: Re: [FSL] melodic segmentation fault > To: [log in to unmask] > > Hi, > > What platform are you running this on? Is this OSX 10.9? > ta > christian > > On 23 Jun 2014, at 15:20, Matt Glasser <[log in to unmask]> wrote: > > > 34013 Segmentation fault: 11 >