Issue No. 07 - July (1998 vol. 47)
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/12.709379
<p><b>Abstract</b>—Cellular automaton (CA) is a promising computer paradigm that can break through the von Neumann bottleneck. Two-dimensional CA is especially suitable for application to pixel-level image processing. Although various architectures have been proposed for processing two-dimensional CA, there are no compact, practical computers. So, in spite of its great potential, CA is not widely used. This paper proposes a highly-parallel two-dimensional cellular automaton architecture called CAM<super>2</super> and presents some evaluation results. CAM<super>2</super> can attain pixel-order parallelism on a single board because it is composed of a CAM, which makes it possible to embed an enormous number processing elements (PEs), corresponding to CA cells, onto one VLSI chip. Multiple-zigzag mapping and dedicated CAM functions enable high-performance CA processing. The performance evaluation results show that 256k CA cells, which correspond to a 512 × 512 picture, can be processed by a CAM<super>2</super> on a single board using deep submicron process technology. The processing speed is more than 10 billion CA cell updates per second. This means that more than a thousand CA-based image processing operations can be done on a 512 × 512 pixel image at video rates (33 msec). CAM<super>2</super> will widen the potentiality of CA and make a significant contribution to the development of compact and high-performance systems.</p>
Cellular automaton, content addressable memory, real-time image processing, SIMD, multiple-zigzag mapping.
T. Ogura and T. Ikenaga, "CAM2: A Highly-Parallel Two-Dimensional Cellular Automaton Architecture," in IEEE Transactions on Computers, vol. 47, no. , pp. 788-801, 1998.