Thanks for your interest.I tried and failed at installing pythran. The instructions are not very easy to follow, specially since I do not have a debian platform
J On 6/25/14, 8:26 AM, serge Guelton wrote:
Dear Jesús Fernandez-Villaverde, After reading your paper ``A Comparison of Programming Languages in Economics'', I wanted to give a try to the Pythran compiler we have been developing for a few years. I had no trouble reproducing your benchmarks. On my laptop: C++ version: 2.17s Numba version: 2.72s To use the pythran compiler, I extracted the ``innerloop`` kernel in a new module and turned it into a native module using Pythran: $> pythran RBC.py $> python RBC_Python_Pythran.py and I get a nice 1.81s, faster than the naïve C++ version \o/ for later reference, I have included the whole setup in attachments. Pythran version is the one from the master at https://github.com/serge-sans-paille/pythran Maybe you could consider add it to your benchmarks?
-- Jesús Fernández-Villaverde Economics-University of Pennsylvania www.econ.upenn.edu/~jesusfv 君臣上下貴賤皆從法，此謂為大治