> If the 1 minute versus 1 second is accurate, then there was a 60x > performance difference in the beginning and Matthew managed to improve > it by a factor of 5 or even slightly over 6 when tuning, so it seems > to me Racket would still be slower by 10x. > >
I checked with the blog author a few days ago, and he reports that one of his test programs runs significantly slower under the development version of Racket vs. 5.2.1, after these changes: cxwangyi <http://cxwangyi.wordpress.com/> says: August 15, 2012 at 7:53 am<http://cxwangyi.wordpress.com/2012/07/29/chinese-whispers-in-racket-and-go/#comment-478> Thanks for the update, and I am now looking forward to v5.3.0.18. One more thing: I just compared v5.2.1 with v5.3.0.17, and found the latter runs times slower. $time racket-5.2.1/bin/racket /tmp/a.rkt real 0m35.430s user 0m26.235s sys 0m5.334s $time ~/racket-5.3.0.17/bin/racket /tmp/a.rkt real 1m54.070s user 0m47.368s sys 0m27.836s Reference: the comments section of http://cxwangyi.wordpress.com/2012/07/29/chinese-whispers-in-racket-and-go/ Maybe someone can follow up with the author directly?
____________________ Racket Users list: http://lists.racket-lang.org/users
