Designing with RoBs for High Performance VLIW Architecture

Abstract

VLIW architecture has become widespread due to the combined bene?ts of simple hardware and compiler extracted instruction level parallelism. The VLIW instruction set architecture and its hardware implementation is tightly coupled and a novel simultaneous multithreading VLIW architecture with dynamic dispatch mechanism which uses RoBs complex logic to maximize ILP has been proposed. Since the resulting dynamic instruction schedule of many applications seldom changes, it is reasonable to store and reuse the schedule instead of reconstructing it each time. The new VLIW architecture shows that it can effectively increase the processor efficiency which improves the performance.