Branch prediction using both global and local branch history information

Meng-Chou Chang, Y. W. Chou

Research output: Contribution to journalArticle

9 Citations (Scopus)

Abstract

As the pipeline depth and issue rate of high-performance superscalar processors increase, the importance of an excellent branch predictor becomes more crucial to delivering the potential performance of a wide-issue, deep pipelined processor. Conventional two-level branch predictors predict the outcome of a branch either based on the local branch history information, comprising the previous outcomes of a single branch, or based on the global branch history information, comprising the previous outcomes of all branches. The authors propose a new branch prediction scheme, called LGshare, which employs both the global and local branch history information simultaneously to improve the branch prediction accuracy for superscalar processors. It is shown that LGshare can achieve higher branch prediction accuracy than conventional two-level predictors when the size of the pattern history table is fixed.

Original languageEnglish
Pages (from-to)33-38
Number of pages6
JournalIEE Proceedings: Computers and Digital Techniques
Volume149
Issue number2
DOIs
Publication statusPublished - 2002 Mar 1

Fingerprint

Branch Prediction
Branch
Superscalar Processor
Predictors
Pipelines
History
Table
High Performance
Predict

All Science Journal Classification (ASJC) codes

  • Theoretical Computer Science
  • Hardware and Architecture
  • Computational Theory and Mathematics

Cite this

@article{5f99a7385a7e4d89b7f3946d1f650648,
title = "Branch prediction using both global and local branch history information",
abstract = "As the pipeline depth and issue rate of high-performance superscalar processors increase, the importance of an excellent branch predictor becomes more crucial to delivering the potential performance of a wide-issue, deep pipelined processor. Conventional two-level branch predictors predict the outcome of a branch either based on the local branch history information, comprising the previous outcomes of a single branch, or based on the global branch history information, comprising the previous outcomes of all branches. The authors propose a new branch prediction scheme, called LGshare, which employs both the global and local branch history information simultaneously to improve the branch prediction accuracy for superscalar processors. It is shown that LGshare can achieve higher branch prediction accuracy than conventional two-level predictors when the size of the pattern history table is fixed.",
author = "Meng-Chou Chang and Chou, {Y. W.}",
year = "2002",
month = "3",
day = "1",
doi = "10.1049/ip-cdt:20020273",
language = "English",
volume = "149",
pages = "33--38",
journal = "IEE Proceedings: Computers and Digital Techniques",
issn = "1350-2387",
publisher = "Institute of Electrical Engineers",
number = "2",

}

Branch prediction using both global and local branch history information. / Chang, Meng-Chou; Chou, Y. W.

In: IEE Proceedings: Computers and Digital Techniques, Vol. 149, No. 2, 01.03.2002, p. 33-38.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Branch prediction using both global and local branch history information

AU - Chang, Meng-Chou

AU - Chou, Y. W.

PY - 2002/3/1

Y1 - 2002/3/1

N2 - As the pipeline depth and issue rate of high-performance superscalar processors increase, the importance of an excellent branch predictor becomes more crucial to delivering the potential performance of a wide-issue, deep pipelined processor. Conventional two-level branch predictors predict the outcome of a branch either based on the local branch history information, comprising the previous outcomes of a single branch, or based on the global branch history information, comprising the previous outcomes of all branches. The authors propose a new branch prediction scheme, called LGshare, which employs both the global and local branch history information simultaneously to improve the branch prediction accuracy for superscalar processors. It is shown that LGshare can achieve higher branch prediction accuracy than conventional two-level predictors when the size of the pattern history table is fixed.

AB - As the pipeline depth and issue rate of high-performance superscalar processors increase, the importance of an excellent branch predictor becomes more crucial to delivering the potential performance of a wide-issue, deep pipelined processor. Conventional two-level branch predictors predict the outcome of a branch either based on the local branch history information, comprising the previous outcomes of a single branch, or based on the global branch history information, comprising the previous outcomes of all branches. The authors propose a new branch prediction scheme, called LGshare, which employs both the global and local branch history information simultaneously to improve the branch prediction accuracy for superscalar processors. It is shown that LGshare can achieve higher branch prediction accuracy than conventional two-level predictors when the size of the pattern history table is fixed.

UR - http://www.scopus.com/inward/record.url?scp=0036506177&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0036506177&partnerID=8YFLogxK

U2 - 10.1049/ip-cdt:20020273

DO - 10.1049/ip-cdt:20020273

M3 - Article

AN - SCOPUS:0036506177

VL - 149

SP - 33

EP - 38

JO - IEE Proceedings: Computers and Digital Techniques

JF - IEE Proceedings: Computers and Digital Techniques

SN - 1350-2387

IS - 2

ER -