[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: efficient kernel or masking algorithm ?

"J.D. Smith" <jdsmith@astro.cornell.edu> writes:
> While I'm on the gripe train, why shouldn't we be able to consistently
> perform operations along any dimension of an array we like with relevant
> IDL routines.  E.g., we can total along a single dimension.  All due
> respect to Craig's CMAPPLY function, but some of these things should be
> much faster.  Resorting to summed logarithms for multiplication is not
> entirely dubious, but why shouldn't we be able to say:

Agree!  Agree!  Agree!  For once we are griping in synchrony :-)

These are exactly the kinds of operations that would be enhanced by
vectorization, but they can't as IDL stands now.

By the way, CMAPPLY doesn't use summed logarithms any more.  It uses
my bestest algorithm that came out of the recent newsgroup discussion.


Craig B. Markwardt, Ph.D.         EMAIL:    craigmnet@cow.physics.wisc.edu
Astrophysics, IDL, Finance, Derivatives | Remove "net" for better response