speed comparison of IDL, numPy, Matlab

John J. Lee phrxy at csv.warwick.ac.uk
Tue Feb 13 08:15:09 CET 2001

On 2001-02-06 15:26:01 PST,  Mark Hadfield (m.hadfield at niwa.cri.nz) wrote:
> 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.

I don't know if this is the kind of thing you'd be looking for, but...

vtk?  I don't know if it does 2D stuff well, though.  OpenDX apparently
does -- 'User Quote' currently on the latter's page:


In a nutshell, I use DX extensively for many types of applications using
both large and small data sets...

I'm basically using it in a 2-d world for remote sensing data
applications. 99% of the people in this area use IDL. DX stands out better
in 3-d applications, but from my experience it is faster and more reliable
than IDL even in 2-d applications (very large data sets). I use DX for
many data types, especially unstructured and polyline and I also use DX as
a debugging tool for just about anything. If an algorithm isn't doing what
I think it should, just start a new DX program from scratch and in minutes
you can be analizing the problem.

                                                             Robert Neely
                                                NASA Langley Research Center


There's PIL too of course, and if you just want to plot graphs, DISLIN,
PGPLOT, Gnuplot (ugly but simple), BLT, Qwt (both GUI toolkits with graph
widgets).  DISLIN has two python interfaces: the standard one and an OO
one that wraps it.

For 3D, there seem to be lots of things with python interfaces -- no idea
what they do, though.


More information about the Python-list mailing list