Oh wow, that's really cool.
On Tue, Dec 21, 2010 at 6:15 PM, Gary Robinson <[email protected]> wrote: >> I thought this email could be relevant for those interested in SciPy / Numpy >> on pypy. With enthought implementing a smaller core and using compatibility >> layers for alternative platforms, it would seem to be a good basis for a >> pypy port. > > All I can do is give a BIG +1 to anything that can get numpy/scipy up more > quickly. > > PyPy is starting to give us the speed we need for statistical/scientific work > on python (the speed it still lacks compared to C or Java is made up for, for > my purposes, by the ease of writing python compared to those languages). The > recent 64-bit functionality lets us process a lot of data. The fast-forward > branch is giving us the multiprocessing we need. (I recognize that there are > other solutions, but for simple things we need to write quickly, the > multiprocessing module is really sweet.) > > The main thing missing now is numpy/scipy. The addition of that will make > PyPy a huge win in the scientific community, IMO. > > Anyway, I mention it in case the opinion of one person who is using Python in > production for statistical processing is of interest. > > -- > > Gary Robinson > CTO > Emergent Discovery, LLC > personal email: [email protected] > work email: [email protected] > Company: http://www.emergentdiscovery.com > Blog: http://www.garyrobinson.net > > > > > _______________________________________________ > [email protected] > http://codespeak.net/mailman/listinfo/pypy-dev > _______________________________________________ [email protected] http://codespeak.net/mailman/listinfo/pypy-dev
