Raymond Toy <[EMAIL PROTECTED]> writes:

>>>>>> "Nicolas" == Nicolas Neuss <[EMAIL PROTECTED]> writes:
>
>     Nicolas> Maybe the bottleneck is mainly the random number generator?  I 
> have written
>     Nicolas> the following routine GASDEV which is about twice as fast as
>     Nicolas> ONE-NORMAL-RAND on my machine (and is hopefully a reasonable 
> replacement to
>     Nicolas> it - note that I am no random number expert).
>
> Also, in clocc (clocc.sourceforge.net), there are implementations of 5
> different Gaussian generators (along with generators for other
> distributions).  Of the 5, Marsaglia's Ziggurat method is far and away
> the fastest.
>
> Ray

Very nice!  It is three times faster than mine, and should therefore be
about six times faster than ONE-NORMAL-RAND.

Thank you for the info (and for writing the routine:-),

Nicolas.


Reply via email to