On 3/1/07, Robert Kern <[EMAIL PROTECTED]> wrote:
Charles R Harris wrote: > Looks like function call overhead has gone way up or the cost of > returning a float vs an array has gone way up. The loop overhead is > about .01 and not significant. So something is definitely wrong here. > Time to go look in trac ;) It might have been when Travis put in the broadcasting capabilities to the nonuniform distribution generators.
Could be, the time is about right: changeset 2892, 7/25/06. Hmmm, there's a lot of code there. Must say nothing jumps out at me 8) Chuck
_______________________________________________ Numpy-discussion mailing list [email protected] http://projects.scipy.org/mailman/listinfo/numpy-discussion
