virtualmeet wrote:
> Sorry to say it again, but I really don't see from where the loss of
> precision for PovRay can came from in case it's using this technic: In the
> deepest level, PovRay is using "only" the CPU "welded" fcts...
> Cheers,
> Taha

OK, let me clarify something.  Are you saying that
a) your procedure will first evaluate a function X number of times and 
store the result, looking and values in this table later and, if it 
can't find them, interpolating values?

or b) your procedure will, as functions are evaluated, store the result 
of those evaluations in a lookup table which is then referenced later in 
case the same function with the same input needs to be called again?

If it's a), then the interpolation is the approximation.  This is the 
method I believed you were espousing, and if I was mistaken then I 
apologize.

If it's b), then I would wonder how often a function will be called with 
input identical to a previous one.  I suppose tests would have to be 
done to determine whether or not this would actually benefit, and those 
would have to be done on an application - by application basis.

...Chambers

Post a reply to this message