Craig DeForest <[email protected]> wrote:

> I am a bit puzzled how these other packages manage to go so much  
> faster...

It's mainly Scilab and Octave that really perform. NumPy seems to
suffer similar problems as PDL, but to a lesser extent.

NumPy's performance is consistently about 33% better than PDL, and your
patch appears to cut that gap in half. On the other hand, the
difference between PDL/NumPy and Scilab/Octave is not consistent. The
difference grows as the matrix gets larger. So clearly Scilab and
Octave are doing something special to scale better.

Daniel.

_______________________________________________
Perldl mailing list
[email protected]
http://mailman.jach.hawaii.edu/mailman/listinfo/perldl

Reply via email to