Re: Implementation issues for fractals ?
Speed isn't everything.
We did a lot at 1 megahertz with a basic interpreter
( not a compiler) in the past...
Content and it's relevance is probably more important.
Now resolution ( pixelation of images) can be important
as it determines how pretty your image is printed out.
For that you do need some speed or the images take days.
None of this is very new...
adam majewski wrote:
anastasiA napisał(a):
Hello.
I've been dabbling around with fractals, more specifically
implementing Mandelbrot and Julia sets, and I was pretty proud of what I
did until I saw some implementations going over 20 times faster than
mine ;-).
Seriously, I would like to have some tips on implementing issues. A lot
of time is spent on squaring, square rooting and calculating complex
numbers and this is probably the most time consuming aspect of the whole
process. I was wondering if anyone would be aware of tricks on
accelarating these processes instead of using the regular C functions.
I've even heard it was possible to get rid of floting point calculations
and use integer mathematics though I don't know how this would affect
precision.
I'm not interested in hacks that fit one program's particularity. More
of a tool box for the aspiring fractal coder to speed up things a bit
(beside shifts and look-up tables: that I learned in game programming
:-))
Hope I'm not asking too much.
Happy new Year to all.
--
a.
http://www.mrob.com/pub/muency/speedimprovements.html
Adam Majewski
republika.pl/fraktal
.
