[Numpy-discussion] How to speed up this function?

Charles R Harris charlesr.harris at gmail.com
Tue Dec 5 12:33:30 CST 2006

On 12/5/06, Charles R Harris <charlesr.harris at gmail.com> wrote:
> On 12/3/06, fsenkel at verizon.net <fsenkel at verizon.net> wrote:
> >
> > Hello,
> >
> > I'm taking a CFD class, one of the codes I wrote runs very slow. When I
> > look at hotshot is says the function below is the problem. Since this is an
> > explicit step, the for loops are only traversed once, so I think it's caused
> > by memory usage, but I'm not sure if it's the local variables or the loop? I
> > can vectorize the inner loop,  would declaring the data structures in the
> > calling routine and passing them in be a better idea than using local
> > storage?
> >
> > I'm new at python and numpy, I need to look at how to get profiling
> > information for the lines within a function.
> >
> >
> > Thank you,
> >
> > Frank
> >
> >
> > PS
> > I tried to post this via google groups, but it didn't seem to go
> > through, sorry if it ends up as multiple postings

> Explicit indexing tends to be very slow. I note what looks to be a lot of
> differencing in the code, so I suspect what you have here is a PDE.  Your
> best bet in the short term is to vectorize as many of these operations as
> possible, but because the expression is so complicated it is a bit of a
> chore to see just how.  It your CFD class allows it, there are probably
> tools in scipy that are adapted to this sort of problem, and in particular
> to CFD. Sandia also puts out PyTrilinos,
> http://software.sandia.gov/trilinos/packages/pytrilinos/, which provides
> interfaces to distributed and parallel PDE solvers. It's big iron software
> for serious problems, so might be a bit of overkill for your applications.

If it is a PDE, you might also want to look into sparse matrices. Other
folks here can tell you more about that.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://projects.scipy.org/pipermail/numpy-discussion/attachments/20061205/a6d33799/attachment.html 

More information about the Numpy-discussion mailing list