Dmitry Olshansky wrote:
> My measurements show that...
> (-O2 gives approx 2 time impovements).
> ...using RandomGen and State monad to generate a list gives at least 4 times
> improvements (on 1 000 000 items).

You earlier said:

> this takes over twice as long as a naively implemented
Python program

So now our "naive" Haskell - ordinary usage of Data.Map
and System.Random, without resorting to things like
unboxed arrays - is beating naive Python handily. Correct?

Or is this with an alternate RNG? Although I think even that
would be fair, since Python uses Mersenne.

Regards,
Yitz
_______________________________________________
Haskell-Cafe mailing list
Haskell-Cafe@haskell.org
http://www.haskell.org/mailman/listinfo/haskell-cafe

Reply via email to