Running out of memory when loading mat-file

Hi,

When I run my model, I save all the contents in a .mat file, which becomes approximately 700mb. Subsequently, when I load it my computer becomes really slow and uses 1.200mb of memory. What can I do? I have R2011a, so I cannot use matfile yet. Furthermore, I've got a 32-bit 2GHz dual core with 2gb RAM. Is there a way to solve this?

Thanks in forward,

Tim

Below is the script that I use to save the .mat file:

   clear all
  use multidimensional_LV (3).ini;
  rand('state', sum(100*clock));
  simtime 0 5000 5000;

   H=[   1     1     1     1     1     1
      0     1     1     1     1     1
      0     0     1     1     1     1
      0     0     0     1     1     1
      0     0     0     0     1     1
      0     0     0     0     1     1
      0     0     0     0     0     1
      0     0     0     0     0     1
      0     0     0     0     0     1
      0     0     0     0     0     1];

    lenZ=[1:6]';                                
    lenA=[1:10]';
    muZ=(lenZ-1)*9/5+1';
    H=normpdf(repmat(lenA,1,6),repmat(muZ',10,1),ones(10,6)*1);

    ncnt=20;
    runs = 100;
    mainloop=5;

    tmpl='NESTNORMPDF_%2.2d';

    for newtry=1:mainloop
        Hs=zeros(ncnt,10,6);
        as=zeros(ncnt,10,10);
        As=zeros(ncnt,runs,10);
        Zs=zeros(ncnt,runs,6);
      stuff=zeros(ncnt,2);
      results=cell(ncnt,runs);

      cnt=1;
      use multidimensional_LV (3).ini;
      simtime 0 5000 5000;

      H=[   1     1     1     1     1     1
          0     1     1     1     1     1
          0     0     1     1     1     1
          0     0     0     1     1     1
          0     0     0     0     1     1
          0     0     0     0     1     1
          0     0     0     0     0     1
          0     0     0     0     0     1
          0     0     0     0     0     1
          0     0     0     0     0     1];

      lenZ=[1:6]';                                %
      lenA=[1:10]';
      muZ=(lenZ-1)*9/5+1';
      H=normpdf(repmat(lenA,1,6),repmat(muZ',10,1),ones(10,6)*1);

      while cnt<=ncnt

          Hs(cnt,:,:)=H;
          as(cnt,:,:)=a;
          savepar(sprintf('ini/run%s-%d',datestr(now(),'yyyymmdd'),cnt))
          newnest=nestedloop2(H);
          [n_group phyto zoo Q]=modularity_bipartite(H);
          stuff(cnt,:)=[newnest;Q];
          tic
          for i=1:runs;

              randinit 5;
              As(cnt,i,:)=A;
              Zs(cnt,i,:)=Z;
              stabil 20000;
              n=nmaxima;

              results{cnt,i}=n;
              progression=[newtry cnt i]          
          end;
          toc
          ouweH=H;
          found=false;
          while ~found
              H=ouweH;
              vindpos=find(H>0.001);
              vindzeros=find(H<0.001);
              vindpos=vindpos(randperm(size(vindpos,1)));
              vindzeros=vindzeros(randperm(size(vindzeros,1)));
              twopos=vindpos(1:2);
              twozeros=vindzeros(1:2);
              helpH=H(twozeros);
              H(twozeros)=H(twopos);
              H(twopos)=helpH
              found=all(sum(H)>0) && all(sum(H,2)>0);
          end

          cnt=cnt+1;

          fnam=sprintf(tmpl,newtry);

          save(fnam,'results','As','Zs','Hs','stuff', 'as');

      end
  end

Answer 1

[Jan]

When the MAT file has 700MB and it is written in a compressed format (-v7.3), loading it will need as much RAM as it needs. Please check "whos" after loading the file to see how much memory the data need.

There is no magic trick to get more memory, but only the trivial suggestions:

• Install more RAM
• Use a 64-bit OS and Matlab