THE VLSI IMPLEMENTATION OF A FINE-GRAINED PARALLEL ARCHITECTURE

Abstract

This thesis evolved out of a fine-grained parallel processor designed by John Cleary. The design was improved to increase the speed and provide testability. The resultant design was then implemented as a VLSI layout. Estimates of power consumption and clocking speeds were then derived from the layout using SPICE simulations. Evaluation shows that the power estimate is within practical limits of chip operation when packaged. Issues of fanout, metal migration, and supply line noise are also addressed.