A Comparison of Superscalar and Decoupled Access/Execute Architectures

Matthew Farrens, Pius Ng, and Phillip Nico

Department of Computer Science
University of California, Davis
Davis, CA 95616

Abstract

This paper presents a comparison between superscalar and decoupled access/execute architectures. Both architectures attempt to exploit instruction-level parallelism by issuing multiple instructions per cycle, employing dynamic scheduling to maximize performance. Simulation results are presented for four different configurations, demonstrating that the architectural queues of the decoupled machines provide similar functionalities as register renaming, dynamic loop unrolling, and out-of-order execution of the superscalar machines with significantly less complexity.

Full paper Appears in Proceeding of the Twenty-sixth Annual International Symposium on Microarchitecture, December 1993.