On Thu, Jul 29, 2010 at 3:41 AM, Sturla Molden <stu...@molden.no> wrote:
> > Thu, 29 Jul 2010 02:40:00 +0200, Sturla Molden wrote: > > > Want to submit a patch? ;) > > I had this in a Matlab MEX file I used for my dissertation. The ziggurat > is not hard to program, even for general PDFs. We could even fill in the > ziggurat table using numerical integration, and code one for general PDFs > (NumPy overhead for the tail fallback could be amortized). There is a > timing here claiming ziggurat with mt19937 is 3.9 times faster than > Box-Müller: > > We have various implementations floating about. There is an adaptation of my ziggurat for parallel mt19937 being done by Charles Brunet and after he finishes testing we can put fit it into numpy. However we need a new name for it, znormal or some such. Suggestions are welcome. Chuck
_______________________________________________ NumPy-Discussion mailing list NumPy-Discussion@scipy.org http://mail.scipy.org/mailman/listinfo/numpy-discussion
