Re: [glob2-devel] To Erik: My Profiling Information

[Leo Wandersleb]

> > Those are the new results. Much different. Perlin and clouds seem to 
> > dominate?

doh! that's me.
the perlin class is copied as is. i should have noted the origin as legal stuff 
should not be taken that easily as i might have taken it back then.
yes i checked the license but maybe should do so again as i derived my "we can 
use it" from some facts that were not "you may use it for any purpose free of 
charge " etc. but rather some "why don't you do it that way: myPerlin.cpp"

Principally I am able to make up my own Perlin class from what i remember from 
my math-studies and yes it would be possible to do in in integer. Maybe time 
allows.

> Is there any chance that the noise can be computed using integers, or 
> cached?

cached? clouds are alpha(x,y,t). all based on a set of some 20B of random data. 
in that this is quite the opposite of cacheing to not have to compute. those 
20B of random data define alpha on an area [inf]^2 for all time past and 
future. don't think that we want to cache that :)

i'm thinking about changing clouds as i would want to have wind effects to 
couple with the new particle engine. nice would be local tourbulences. wind now 
is global. i can't see yet that the perlin approach would be able to fit in 
here but again i would want to define some alpha(x,y,t).

Thanx for your great work.

Leo Wandersleb