Hi folks,
On Tue, Aug 25, 2009 at 12:08 PM, Kurt Smith<[email protected]> wrote:
> Weave-like support in Cython -- Fernando Perez (not at the BoF)
> suggested that Cython could assume the same functionality of
> scipy.weave (see: http://www.scipy.org/Weave), since scipy.weave is
> poorly documented and could use more maintenance than it is currently
> getting. I'm personally very interested in this -- it seems that
> there is much overlap between scipy.weave and the combination of
> pyximport & the pure-python mode of Cython. It seems that with a bit
> of work from an interested user and some guidance from yours truly, we
> could improve pyximport & Cython's pure-python mode to incorporate the
> good stuff from scipy.weave, but better ;-) I intend to write-up a
> CEP with my thoughts sometime in the near future.
>
Kurt, thanks for resurrecting this comment. Indeed, weave is an
extremely useful piece of functionality but one that has been somewhat
neglected for years. There's a lot of good in there though, and the
need for it is significant. It's *really* cool to do things like:
def prod(m, v):
"""Matrix-vector multiply via weave/blitz++"""
nrows, ncolumns = m.shape
assert v.ndim==1 and ncolumns==v.shape[0],"Shape mismatch in prod"
res = np.zeros(nrows, float)
code = r"""
for (int i=0; i<nrows; i++)
{
for (int j=0; j<ncolumns; j++)
{
res(i) += m(i,j)*v(j);
}
}
"""
err = inline(code,['nrows', 'ncolumns', 'res', 'm', 'v'], verbose=2,
type_converters=converters.blitz)
return res
This makes interactive experimentation and the writing of
loop/indexing-intensive code really easy. But by itself, weave is in
serious danger of bitrot, and right now Cython has the momentum and
concentrated effort for this problem domain.
Everyone is extremely impressed with the amazing work you've all done,
and rooting for cython to continue to improve, as you are solving a
number of key problems to make python a really complete platform for
scientific computing. It's been lots of fun to see cython steadily
attack a number of problems that have been a real issue for years.
So if Cython absorbs/extends/integrates/develops/whatever the good
ideas and code from weave, so that this type of workflow is also
supported, it would be great. I am really happy to see f2py grow
into fwrap, and the weave machinery (or at least parts of it) seems
like a logical addition to the cython-based toolkit.
One particular use case for which weave.inline was invaluable to me in
the past was the generation of fast looping code for various
dimensions. I had to write function evaluation codes where the user
could supply a C expression (as a string, that would #define NDIM to
indicate the dimensionality of the problem). I could then easily
generate a few methods for D=1..6 (the range I needed to cover) that
would do certain key operations very fast. The bulk of the code used
numpy arrays to be dimension agnostic, but a few methods really needed
explicit loops and with weave, it was made completely transparent to
the user, all of it being done at runtime (for us, needing gcc at
runtime was OK). I mention this to give you some feedback on the use
cases for weave (there are others), but I hope that in the end, all of
this class of functionality will be available through a unified entry
point, and if that's cython, all the better.
Thanks again to the team for your great work!
Cheers,
f
_______________________________________________
Cython-dev mailing list
[email protected]
http://codespeak.net/mailman/listinfo/cython-dev
