[Octave-bug-tracker] [bug #29768] logm fix

[Rik]

Follow-up Comment #11, bug #29768 (project octave):

The proposed new algorithm is awfully slow compared to the old one.  I ran a
test with the new algorithm, including new sqrtm, on a random 500x500 matrix. 
The results were 13.9 seconds for the new calculation versus 2.8 seconds for
the old calculation (5x slower).  Numerically, the two results were within
rounding errors of each other.

If the only problem with the old calculation is the occasional problem of
inverting the eigenvector matrix, wouldn't it make sense to use the Pade
approach only for these matrices and use the fast eigenvalue decomposition for
the rest?

In pseudo-code

[V, D] = eig(a);
if (condest (V) < _SINGULAR_MATRIX_CUTOFF_)
  Eigenvalue approach
else  
  Pade approach
endif

    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/bugs/?29768>

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