The java applet above simulates the signal propagation that comprises all communication between processors in the DES Cracker algorithm described in the paper referenced below. This page discusses only the signals themselves; a more complete description of the problem and the cell-matrix-based solution we designed is provided in the paper.
     If you don't see anything above this block of text, check your browser's settings and make sure that Java is enabled.
     Click here to get a description of the signals and passing schemes, and a guide to what you're seeing above.
     The java simulation responds to mouse clicks. You can click on a square to make it the matching processor. Then hit the Restart Simulation button. The next simulation will show the propagation of the match signal from that square rather than the previous one.
More information on the DES Cracker (algorithm, processor setup, hardware, and performance numbers) is available in the paper itself. This algorithm is a good example of what we call spatially distributed algorithms, ones in which the problem is described such that it can be farmed out over many, many similar processors with minimal interprocessor communication. It is a time-space tradeoff, in which a lot of hardware is used to solve the problem in constant O(1) time. This type of algorithm assumes computing hardware that can achieve a MISD (multiple instruction, single data) architecture. Cell matrices can perform this class of MISD architecture quite well.
    You can also see a more detailed version, of signals between and inside processors, in the 4 bit version of this DES Cracker running on cell matrix simulators.
"The Cell Matrix: An Architecture for Nanocomputing," by L. Durbeck and N. Macias, 2000. Available in several formats from our publications list.