Abstract
A parallel processor specially designed for an important problem in theoretical physics is described. The final device will contain 256 nodes running in lock-step in a SIMD mode with a computational power of 4 billion 22-bit floating point operations per second. Each node is controlled by an Intel 80286/287 microprocessor, contains 160K bits of memory and has a pipelined, microprogrammable arithmetic unit which performs floating point multiplications and additions. The nodes are interconnected in a two-dimensional rectangular grid so that vectorized arithmetic can be performed using data stored on any two adjacent nodes.

This publication has 9 references indexed in Scilit: