Combining local and global history hashing in perceptron branch prediction

C. Y. Ho; Anthony S. S. Fong

doi:10.1109/ICIS.2007.81

Combining local and global history hashing in perceptron branch prediction

C. Y. Ho
, Anthony S. S. Fong

Department of Electrical Engineering

Research output: Chapters, Conference Papers, Creative and Literary Works › RGC 32 - Refereed conference paper (with host publication) › peer-review

6 Citations (Scopus)

Abstract

As the instruction issue rate and depth of pipelining increase, branch prediction is considered as a performance hurdle for modern processors. Extremely high branch prediction accuracy is essential to deliver their potential performance. Many perceptron branch predictors have been investigated to improve the dynamic branch prediction in recent years. This paper introduces combining local history hashing and global history hashing in perceptron branch prediction. This proposed perceptron predictor utilizes self-history as well as global history in indexing different weights of a perceptron. The simulation results show that our proposed perceptron predictor is more accurate than the one using either global history hashing or local history hashing alone. Our proposed perceptron predictor is able to achieve 4.13% misprediction rate and even 0.45% misprediction rate in some cases. And it has an improvement of 9.21% over using global history hashing alone, the mapping scheme proposed by Tarjan and Skadron. © 2007 IEEE.

Original language	English
Title of host publication	Proceedings - 6th IEEE/ACIS International Conference on Computer and Information Science (ICIS 2007) In conjunction with 1st IEEE/ACIS International Workshop on e-Activity (IWEA 2007)
Editors	Roger Lee, Morshed U. Chowdhury, Sid Ray, Thuy Lee
Publisher	IEEE
Pages	54-59
ISBN (Print)	0769528414, 9780769528410
DOIs	https://doi.org/10.1109/ICIS.2007.81
Publication status	Published - 2007
Event	6th IEEE/ACIS International Conference on Computer and Information Science (ICIS 2007) In conjunction with 1st IEEE/ACIS International Workshop on e-Activity (IWEA 2007) - Melbourne, Australia Duration: 11 Jul 2007 → 13 Jul 2007

Conference

Conference	6th IEEE/ACIS International Conference on Computer and Information Science (ICIS 2007) In conjunction with 1st IEEE/ACIS International Workshop on e-Activity (IWEA 2007)
Place	Australia
City	Melbourne
Period	11/07/07 → 13/07/07

Funding

The work described in this paper was partially supported by the City University of Hong Kong, Strategic Research Grant 7001847.

Access Output

10.1109/ICIS.2007.81

Other Access Options

Check CityUHK Library

Permanent Link

Right click to copy link

Cite this

Ho, C. Y., & Fong, A. S. S. (2007). Combining local and global history hashing in perceptron branch prediction. In R. Lee, M. U. Chowdhury, S. Ray, & T. Lee (Eds.), Proceedings - 6th IEEE/ACIS International Conference on Computer and Information Science (ICIS 2007) In conjunction with 1st IEEE/ACIS International Workshop on e-Activity (IWEA 2007) (pp. 54-59). IEEE. https://doi.org/10.1109/ICIS.2007.81

Ho, C. Y. ; Fong, Anthony S. S. / Combining local and global history hashing in perceptron branch prediction. Proceedings - 6th IEEE/ACIS International Conference on Computer and Information Science (ICIS 2007) In conjunction with 1st IEEE/ACIS International Workshop on e-Activity (IWEA 2007). editor / Roger Lee ; Morshed U. Chowdhury ; Sid Ray ; Thuy Lee. IEEE, 2007. pp. 54-59

@inproceedings{0d34eea265fc4cbb82e910c4fb9fd4e2,

title = "Combining local and global history hashing in perceptron branch prediction",

abstract = "As the instruction issue rate and depth of pipelining increase, branch prediction is considered as a performance hurdle for modern processors. Extremely high branch prediction accuracy is essential to deliver their potential performance. Many perceptron branch predictors have been investigated to improve the dynamic branch prediction in recent years. This paper introduces combining local history hashing and global history hashing in perceptron branch prediction. This proposed perceptron predictor utilizes self-history as well as global history in indexing different weights of a perceptron. The simulation results show that our proposed perceptron predictor is more accurate than the one using either global history hashing or local history hashing alone. Our proposed perceptron predictor is able to achieve 4.13\% misprediction rate and even 0.45\% misprediction rate in some cases. And it has an improvement of 9.21\% over using global history hashing alone, the mapping scheme proposed by Tarjan and Skadron. {\textcopyright} 2007 IEEE.",

author = "Ho, \{C. Y.\} and Fong, \{Anthony S. S.\}",

year = "2007",

doi = "10.1109/ICIS.2007.81",

language = "English",

isbn = "0769528414",

pages = "54--59",

editor = "Roger Lee and Chowdhury, \{Morshed U.\} and Sid Ray and Thuy Lee",

booktitle = "Proceedings - 6th IEEE/ACIS International Conference on Computer and Information Science (ICIS 2007) In conjunction with 1st IEEE/ACIS International Workshop on e-Activity (IWEA 2007)",

publisher = "IEEE",

address = "United States",

note = "6th IEEE/ACIS International Conference on Computer and Information Science (ICIS 2007) In conjunction with 1st IEEE/ACIS International Workshop on e-Activity (IWEA 2007) ; Conference date: 11-07-2007 Through 13-07-2007",

}

Ho, CY & Fong, ASS 2007, Combining local and global history hashing in perceptron branch prediction. in R Lee, MU Chowdhury, S Ray & T Lee (eds), Proceedings - 6th IEEE/ACIS International Conference on Computer and Information Science (ICIS 2007) In conjunction with 1st IEEE/ACIS International Workshop on e-Activity (IWEA 2007). IEEE, pp. 54-59, 6th IEEE/ACIS International Conference on Computer and Information Science (ICIS 2007) In conjunction with 1st IEEE/ACIS International Workshop on e-Activity (IWEA 2007), Melbourne, Victoria, Australia, 11/07/07. https://doi.org/10.1109/ICIS.2007.81

Combining local and global history hashing in perceptron branch prediction. / Ho, C. Y.; Fong, Anthony S. S.
Proceedings - 6th IEEE/ACIS International Conference on Computer and Information Science (ICIS 2007) In conjunction with 1st IEEE/ACIS International Workshop on e-Activity (IWEA 2007). ed. / Roger Lee; Morshed U. Chowdhury; Sid Ray; Thuy Lee. IEEE, 2007. p. 54-59.

Research output: Chapters, Conference Papers, Creative and Literary Works › RGC 32 - Refereed conference paper (with host publication) › peer-review

TY - GEN

T1 - Combining local and global history hashing in perceptron branch prediction

AU - Ho, C. Y.

AU - Fong, Anthony S. S.

PY - 2007

Y1 - 2007

N2 - As the instruction issue rate and depth of pipelining increase, branch prediction is considered as a performance hurdle for modern processors. Extremely high branch prediction accuracy is essential to deliver their potential performance. Many perceptron branch predictors have been investigated to improve the dynamic branch prediction in recent years. This paper introduces combining local history hashing and global history hashing in perceptron branch prediction. This proposed perceptron predictor utilizes self-history as well as global history in indexing different weights of a perceptron. The simulation results show that our proposed perceptron predictor is more accurate than the one using either global history hashing or local history hashing alone. Our proposed perceptron predictor is able to achieve 4.13% misprediction rate and even 0.45% misprediction rate in some cases. And it has an improvement of 9.21% over using global history hashing alone, the mapping scheme proposed by Tarjan and Skadron. © 2007 IEEE.

AB - As the instruction issue rate and depth of pipelining increase, branch prediction is considered as a performance hurdle for modern processors. Extremely high branch prediction accuracy is essential to deliver their potential performance. Many perceptron branch predictors have been investigated to improve the dynamic branch prediction in recent years. This paper introduces combining local history hashing and global history hashing in perceptron branch prediction. This proposed perceptron predictor utilizes self-history as well as global history in indexing different weights of a perceptron. The simulation results show that our proposed perceptron predictor is more accurate than the one using either global history hashing or local history hashing alone. Our proposed perceptron predictor is able to achieve 4.13% misprediction rate and even 0.45% misprediction rate in some cases. And it has an improvement of 9.21% over using global history hashing alone, the mapping scheme proposed by Tarjan and Skadron. © 2007 IEEE.

UR - https://www.scopus.com/pages/publications/46749121780

UR - https://www.scopus.com/record/pubmetrics.uri?eid=2-s2.0-46749121780&origin=recordpage

U2 - 10.1109/ICIS.2007.81

DO - 10.1109/ICIS.2007.81

M3 - RGC 32 - Refereed conference paper (with host publication)

SN - 0769528414

SN - 9780769528410

SP - 54

EP - 59

BT - Proceedings - 6th IEEE/ACIS International Conference on Computer and Information Science (ICIS 2007) In conjunction with 1st IEEE/ACIS International Workshop on e-Activity (IWEA 2007)

A2 - Lee, Roger

A2 - Chowdhury, Morshed U.

A2 - Ray, Sid

A2 - Lee, Thuy

PB - IEEE

T2 - 6th IEEE/ACIS International Conference on Computer and Information Science (ICIS 2007) In conjunction with 1st IEEE/ACIS International Workshop on e-Activity (IWEA 2007)

Y2 - 11 July 2007 through 13 July 2007

ER -

Ho CY, Fong ASS. Combining local and global history hashing in perceptron branch prediction. In Lee R, Chowdhury MU, Ray S, Lee T, editors, Proceedings - 6th IEEE/ACIS International Conference on Computer and Information Science (ICIS 2007) In conjunction with 1st IEEE/ACIS International Workshop on e-Activity (IWEA 2007). IEEE. 2007. p. 54-59 Epub 2007 Jul 23. doi: 10.1109/ICIS.2007.81

Combining local and global history hashing in perceptron branch prediction

Abstract

Conference

Funding

Access Output

Other Access Options

Permanent Link

Other Files and Links

Fingerprint

Cite this