speed comparison of IDL, numPy, Matlab

Mark Hadfield m.hadfield at niwa.cri.nz
Wed Feb 7 00:10:55 CET 2001


"Benyang Tang" <btang at pacific.jpl.nasa.gov> wrote in message
news:3A7EEE5C.6419B707 at pacific.jpl.nasa.gov...
>
> Out of curiosity, I did a quick benchmark test of IDL, NumPy and Matlab on
my
> desktop machine. I know benchmarking is a complicated issue; don't take my
> naive test too serious.

I don't take it seriously at all, because you've only exercised a tiny
fraction of the code in each product.

As another very small contribution to the IDL vs NumPy vs Matlab discussion,
may I offer my observation that, when you are forced to write (quelle
horreur!) a for loop, IDL will process that loop significantly faster than
Matlab and orders of magnitude faster than Python. (I'm afraid I can't offer
the numbers to back up that assertion right now--it's based on comparisons I
did some time ago.)

Speaking as a long-time IDL user who has dabbled with Matlab & Python/NumPy,
I'd love to switch to NumPy, and I'd do it tomorrow if the graphics
capabilities were there.

---
Mark Hadfield
m.hadfield at niwa.cri.nz  http://katipo.niwa.cri.nz/~hadfield/
National Institute for Water and Atmospheric Research
PO Box 14-901, Wellington, New Zealand






More information about the Python-list mailing list