« Home

Further improvements of the computational model:

First of all i'm a complete newbie to java. The basic structure i took from an applet showing fractals.
I would like to increase the speed of calculation i.e. when the no. of dimensions is increased, the gaps between the no. of points generating semi-regular convex polytopes will be higher and the time to generate a stable structure will exponentially increase.
Currently i am doing a giant search for semi-regulars beyond 10D within scope [20][1000]. Even with the faster converging 1/R^(n-1) law, suggested by Harmen de Boer from the Netherlands, this will take an estimated 3 months of computer time on a normal modern PC.
Ideally the model should be implemented as distributed via internet where each computer connected will calculate a small portion together forming a peer to peer supercomputer. I am seeking help to set up such (for examples see http://www.ii.uib.no/grisk/ and http://www.distributed.net/history.html.en initiatives.)
Further improvements are adding features like rotating the graphical output, faster search of simplices and cross-polytopes, importing and exporting structures and improve algorithms in general.
Furthermore some bugs, which were introduced when porting the model to java, need to be solved;
e.g. when incrementing/decrementing points and/or dimensions calculation of the iterations is not completed and sometimes causes hang-ups and irregularities.
People who like the model and know about Java are invited to load the model locally and improve it.

For any remarks you can contact me at symen@2on.com.

2004, Symen H. Hovinga

« Home