On Tue, 14 Dec 2004, Raymond Toy wrote:

> On x86, the main part of the random number generator is in hand-coded 
> assembly.  I'm pretty sure the part that converts them to double-floats 
> is also pretty efficient.

Good to know. But anyway, with random numbers there is always some 
trade-off between quality and speed. So, for problems that either need a 
very fast RNG that may well be primitive, or for problems where 
cryptographic strength is important, the internal RNG might be 
inappropriate, I suppose.

> > More or less. At least with (complex double-float) you may also run into 
> > trouble if you run out of FPU registers. Then again, the system will 
> > cons numbers.
> 
> As I understand it, this doesn't happen.  When you run out of registers, 
> the numbers are placed on the number stack, which doesn't cons.  But I 
> might be wrong about this.

I think to remember I experienced just this problem a few years ago, but 
I'd have to check again to really make sure.

-- 
regards,               [EMAIL PROTECTED]              (o_
 Thomas Fischbacher -  http://www.cip.physik.uni-muenchen.de/~tf  //\
(lambda (n) ((lambda (p q r) (p p q r)) (lambda (g x y)           V_/_
(if (= x 0) y (g g (- x 1) (* x y)))) n 1))                  (Debian GNU)


Reply via email to