Rapid System Prototyping for High Performance Reconfigurable Computing

This paper proposes a soft-reconfigurable coarse grain platform, called J-platform, based on a very few types of computational cells, and shows how a range of advanced applications can be mapped for achieving high performance. Two very flexible cells are proposed that account for the versatility of the approach. These are the MA_PLUS, which is an enhanced Multiply-Add cell, and the UNL, a Universal NonLinear cell. These two cells, together, provide an unprecedented capability for coarse-grain reconfigurable computing, as discussed in the paper. Although the applications of the platform range from FIR filtering of images to large-scale inverse problems, the paper demonstrates mapping of two specific advanced problems, namely (1) Reconstruction of Computerized Tomography images from fan beam projections, and (2) Color Conversion of Video from the RGB to HSI domains. Speed-up by factors of 20 or more over today's work stations is estimated.