Why running out of memory here?

[Jeff]

Could someone explain why the following code works fine, but if I uncomment any one of the fflush(stdout()) lines, then it tells me "error: memory exhausted or requested size too large for range of Octave's index type -- trying to return to prompt". The files being read are 100,000 by 64 floating points numbers each.

%% Compare cvode solutions after rearranging RHS function.

clear;

p=5;

N=2^p;

srcNew   = '/raid/fpuData/';      % OCTAVE

srcOld   = '/raid/fpuData/bak/';  % OCTAVE

tgt   = '/raid/fpuData/';  % OCTAVE

ftype = '_forward.U.csv';                 % OCTAVE

tmplt='N=%04d_Beta=%05.2f_tol=1e-08_t=00000';       % OCTAVE  % 1-dimension

betas = 4; 12:-1:4;

for beta = betas

  fn=sprintf(tmplt, N, beta);       % 1-dimension

  tic;

  fprintf('Loading %s...\n', [srcNew fn ftype]);

  %fflush(stdout());

  uNew = csvread([srcNew fn ftype]); toc;

  fprintf('Loading %s...\n', [srcOld fn ftype]); 

  %fflush(stdout());

  uOld = csvread([srcOld fn ftype]); toc;

  deltaU = abs(uNew - uOld)/uOld;

  fn = sprintf('N=%04d_Beta=%02.0f_delta.csv', N, beta)

  fprintf('Saving %s...\n', [tgt fn]); 

  %fflush(stdout());

  csvwrite([tgt fn], deltaU); toc;

end