Hardware for cellular automata


The most suitable hardware currently available for implementing cellular automata in is probably a type of electronic circuit called a "cellular-style" FPGA (Field Programmable Gate Array).

Some future technologies will undoubtedly have an impact on the type of physical substrate most suitable for implementing universal cellular automata in:

  • Nanotechnology
    • Nanotechnology will almost certainly result in a break from transmitting and storing information using electricity. Instead, mechanical methods will be used to store and manipulate information on the atomic or molecular levels.
    • Biological computation
      • Biological computation may be seen as essentially a simplified method of approaching nanotechnology. Rather than engineer novel structures from scratch, biological computational approaches hope to make use of modified existing organisms, their products, or their components to provide high tech, nanotechnological machinery for computation. There have already been some minor successes from this field - notably ones using nucleic acids as information storage devices.
    • Crystal computation
      • Crystal computation is one of the hopes for offerring a short-cut to nanotechnological computation in cellular automata. Crystalline structures tesselate in ways which reflect the structure of various types of cellular automata, may be relatively simply constructed (indeed they exhibit strong tendencies towards self-assembly), and have the potential to be a virtually ideal medium for hardware implementation of cellular automata.

  • Quantum computation
    • When it becoms possible, and if it becomes practical, quantum computation may offer the possibility of massively parallelising certain types of operation, in effect by executing them simultaneously in parallel universes.

      There are currently some uncertainties associated with to what degree it will be possible to insulate the computational systems from interactions with the system observing it - any such interactions have the capacity to destroy the details of the computation.

      It has been argued that gravitational processes will cause this type of interaction - regardless of any efforts by experimenters to prevent them. Although gravitation does not prevent interference effects between photons from occurring, quantum computations will necessarily be larger in size and require a number of interactions to occur in serial before they can be used for non-trivial purposes.

      It seems likely that there will be upper limits, both on the size of quantum circuitry and the and time available for performing computations in before the system decoheres. Exactly where these limits lie in practice is not yet known.

      Quantum computation offers perhaps the most in the way of possible benefits, but may be the most distant, and technologically challanging of the techniques listed here.

Tim Tyler | tim@tt1.org | http://cell-auto.com/