This project is read-only.


Computation of next state is too slow, especially when field size is large


Of course this is relative.... but even on my fast laptop, a field size of 1000x800 takes quite some time to compute -- in subsecond range, but need much faster.

SARCASim, an old implementation of CA for Win32 computes and displays MUCH faster.

May need to use unmanaged code to speed things up.

Also, need to consider changing the storage model.

Please read carefully the 4 DOCuments found in this archive.
The author discusses algorithm optimizations. A very inspiring and informative read.


onty wrote Feb 12, 2014 at 1:37 PM

This has been addressed in build 0.10 via:
  1. support for multiple CPUs/cores
  2. making available an optimized algorithms for the Conway transformation
Improving performance is always an ongoing initiative, and so closing this issue does not mean we have achieved perfection (yet!).