Re: [Blitz-support] Periodic boundary conditions?

[Norbert N. <Norbert@Nemec-online.de>]

On Thursday, 9. May 2002 15:31, you wrote:
> > Now, since in the full theory, of course, spacetime
> > is infinite, a way has to be found to handle the
> > boundaries of the finite-sized grid.  The usual way
> > is to introduce periodic boundary conditions, meaning
> > that the site [x,y,z,t] would be identified with
> > [x+N,y,z,t] (in every direction, with N being the
> > size of the grid)
>
> Do you simply want to indroduce a modulus operator into the indices?  I'm
> not sure about inheriting from an Array, but you could certainly define
> your own InifiniteArray class with all the desired functions and operators,
> then just introduce the % operator into the index arguments.

Do you believe that it'll be that simple? I haven't gone through all the 
relavant parts of the library yet, but somehow, I have the feeling that this 
change might interfere with assumptions made at many more parts of the 
library. Well, I guess, the best way to find out is to try and to debug...

B.t.w.: The % operator will only do part of the job - unfortunately, it maps 
negative numbers to negative values. But I fear that might be the least of 
the problems I'll have to face...

I'll keep you informed.

Ciao,
Nobbi

-- 
-- __________________________________________________________
-- JESUS CHRIST IS LORD!
--              To Him, even that machine here has to obey...
--
-- _____________________________________Norbert "Nobbi" Nemec
-- Hindenburgstr. 44   ...   D-91054 Erlangen   ...   Germany
-- eMail: <Norbert@Nemec-online.de>  Tel: +49-(0)-9131-204180
_______________________________________________
Blitz-support mailing list
Bli...@oo...
http://www.oonumerics.org/mailman/listinfo.cgi/blitz-support