Comparison of Memory Access between Superscalar and Vector Processing

YASUDA Yoshiko, TANAKA Teruo, INAGAMI Yasuhiro

Bibliographic Information

Other Title

スーパースカラ方式とベクトル処理方式の比較 : 主記憶アクセス特性に着目して

Description

We make a comparative evaluation or the memory accesses or vector and superscalar processor whose memory accesses are pipelined. In this evaluation, we assume that the configuration or arithmetic units and main memory are the same, and used the most suitable code for each architecture by applying optimizations like loop unrolling. We have round that vector processor achieves higher performance in multiply-add accumulation because or sequential memory access. On the other hand, for inner product, both processors perform stride accesses. Despite the disadvantage or the superscalar processor on multiply-add accumulation, given that both architectures yield similar performances, we conclude that the superscalar processor is general purpose.

Journal

IPSJ SIG Notes

IPSJ SIG Notes 113 233-240, 1995-08-23

Information Processing Society of Japan (IPSJ)

Details 詳細情報について

CRID: 1572543026946882176

NII Article ID: 110002775436

NII Book ID: AN10096105

Text Lang: ja

Data Source

CiNii Articles

Export

Report a problem