[Octave-bug-tracker] [bug #43519] eigs really slow compared to Matlab

[Rik]

Follow-up Comment #7, bug #43519 (project octave):

Summarizing, I was not able to reproduce this bug and jwe's results also show
that eigs on a sparse matrix is faster.  There is probably something else
going on, like swapping, that is particular to the bug reporter's machine.

Following up on comment #6, I have re-written the benchmark to use cputime. 
It would be useful if the bug reporter could run this code and report the
results back here.

Code to cut/paste and execute


function doit (L)
  a = rand (L) > 0.95;
  b = a+a'+0.001*eye (L);
  c = sparse (b);
  t0 = cputime ();
  e = eigs (b);
  tfull = cputime () - t0;
  t0 = cputime ();
  e = eigs (c);
  tsparse = cputime () - t0;
  printf ("L = %d\n", L);
  printf ("full   : %f\n", tfull);
  printf ("sparse : %f\n", tsparse);
endfunction

for L = 1:6
  doit (L*1000)
endfor


Results on my machine:


L = 1000
full   : 0.616038
sparse : 0.120007
L = 2000
full   : 3.868242
sparse : 0.920057
L = 3000
full   : 8.168510
sparse : 1.880118
L = 4000
full   : 16.001000
sparse : 3.780236
L = 5000
full   : 27.577724
sparse : 7.216451
L = 6000
full   : 52.507281
sparse : 11.976749



    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/bugs/?43519>

_______________________________________________
  Message sent via/by Savannah
  http://savannah.gnu.org/