On Jun 13, 1:12 pm, jzakiya <[EMAIL PROTECTED]> wrote: > The paper presents benchmarks with Ruby 1.9.0-1 (YARV). I would love > to see my various prime generators benchmarked with optimized > implementations in other languages. I'm hoping Python gurus will do > better than I, though the methodology is very very simple, since all I > do is additions, multiplications, and array reads/writes.
After playing a little with it, I managed to get a 32-47% improvement on average for the pure Python version, and a 230-650% improvement with an extra "import psyco; psyco.full()" (pasted at http://codepad.org/C2nQ8syr) The changes are: - Replaced range() with xrange() - Replaced x**2 with x*x - Replaced (a,b) = (c,d) with a=c; b=d - Replaced generator expressions with list comprehensions. This was the big one for letting psyco do its magic. I also tried adding type declarations and running it through Cython but the improvement was much less impressive than Psyco. I'm not a Pyrex/Cython expert though so I may have missed something obvious. George -- http://mail.python.org/mailman/listinfo/python-list
