The impact of the register update unit size on multipath execution

Chao-Chin Wu, Kuan Chou Lai, En Hao Liu, Jin Yuan Chen

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Citation (Scopus)

Abstract

Branch prediction is a key mechanism to boost the system performance of a superscalar processor. Though the prediction accuracy rate becomes higher and higher, the mispredicitons still lead to significant performance losses in a wide-issue deep-pipelined superscalar. To address the problem, the technique of multipath execution has been proposed previously, which is capable of executing both paths whenever a lower-confidence conditional branch is encountered. However, because the instructions from different paths share a single register update unit (RUU), they are interleaved in the RUU. In consequence, when a conditional branch is resolved and the instructions on the wrong paths are squashed, all the entries in the resulting holes cannot be reused until they are reclaimed at the commit stage. Since the RUU size is crucial to the performance, it is interesting to know how much can we speedup the performance if the squashed RUU entries can be reused as soon as possible. We have proposed a simple mechanism with very limited hardware resources to achieve this goal. Finally the preliminary simulation results are presented.

Original languageEnglish
Title of host publication2007 IEEE Pacific Rim Conference on Communications, Computers and Signal Processing, Conference Proceedings, PACRIM
Pages113-116
Number of pages4
DOIs
Publication statusPublished - 2007 Dec 1
Event2007 IEEE Pacific Rim Conference on Communications, Computers and Signal Processing, PACRIM - Victoria, BC, Canada
Duration: 2007 Aug 222007 Aug 24

Publication series

NameIEEE Pacific RIM Conference on Communications, Computers, and Signal Processing - Proceedings

Other

Other2007 IEEE Pacific Rim Conference on Communications, Computers and Signal Processing, PACRIM
CountryCanada
CityVictoria, BC
Period07-08-2207-08-24

Fingerprint

Hardware

All Science Journal Classification (ASJC) codes

  • Signal Processing
  • Computer Networks and Communications

Cite this

Wu, C-C., Lai, K. C., Liu, E. H., & Chen, J. Y. (2007). The impact of the register update unit size on multipath execution. In 2007 IEEE Pacific Rim Conference on Communications, Computers and Signal Processing, Conference Proceedings, PACRIM (pp. 113-116). [4313190] (IEEE Pacific RIM Conference on Communications, Computers, and Signal Processing - Proceedings). https://doi.org/10.1109/PACRIM.2007.4313190
Wu, Chao-Chin ; Lai, Kuan Chou ; Liu, En Hao ; Chen, Jin Yuan. / The impact of the register update unit size on multipath execution. 2007 IEEE Pacific Rim Conference on Communications, Computers and Signal Processing, Conference Proceedings, PACRIM. 2007. pp. 113-116 (IEEE Pacific RIM Conference on Communications, Computers, and Signal Processing - Proceedings).
@inproceedings{08ce0ef351f4491780ba91af3561a92c,
title = "The impact of the register update unit size on multipath execution",
abstract = "Branch prediction is a key mechanism to boost the system performance of a superscalar processor. Though the prediction accuracy rate becomes higher and higher, the mispredicitons still lead to significant performance losses in a wide-issue deep-pipelined superscalar. To address the problem, the technique of multipath execution has been proposed previously, which is capable of executing both paths whenever a lower-confidence conditional branch is encountered. However, because the instructions from different paths share a single register update unit (RUU), they are interleaved in the RUU. In consequence, when a conditional branch is resolved and the instructions on the wrong paths are squashed, all the entries in the resulting holes cannot be reused until they are reclaimed at the commit stage. Since the RUU size is crucial to the performance, it is interesting to know how much can we speedup the performance if the squashed RUU entries can be reused as soon as possible. We have proposed a simple mechanism with very limited hardware resources to achieve this goal. Finally the preliminary simulation results are presented.",
author = "Chao-Chin Wu and Lai, {Kuan Chou} and Liu, {En Hao} and Chen, {Jin Yuan}",
year = "2007",
month = "12",
day = "1",
doi = "10.1109/PACRIM.2007.4313190",
language = "English",
isbn = "1424411904",
series = "IEEE Pacific RIM Conference on Communications, Computers, and Signal Processing - Proceedings",
pages = "113--116",
booktitle = "2007 IEEE Pacific Rim Conference on Communications, Computers and Signal Processing, Conference Proceedings, PACRIM",

}

Wu, C-C, Lai, KC, Liu, EH & Chen, JY 2007, The impact of the register update unit size on multipath execution. in 2007 IEEE Pacific Rim Conference on Communications, Computers and Signal Processing, Conference Proceedings, PACRIM., 4313190, IEEE Pacific RIM Conference on Communications, Computers, and Signal Processing - Proceedings, pp. 113-116, 2007 IEEE Pacific Rim Conference on Communications, Computers and Signal Processing, PACRIM, Victoria, BC, Canada, 07-08-22. https://doi.org/10.1109/PACRIM.2007.4313190

The impact of the register update unit size on multipath execution. / Wu, Chao-Chin; Lai, Kuan Chou; Liu, En Hao; Chen, Jin Yuan.

2007 IEEE Pacific Rim Conference on Communications, Computers and Signal Processing, Conference Proceedings, PACRIM. 2007. p. 113-116 4313190 (IEEE Pacific RIM Conference on Communications, Computers, and Signal Processing - Proceedings).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

TY - GEN

T1 - The impact of the register update unit size on multipath execution

AU - Wu, Chao-Chin

AU - Lai, Kuan Chou

AU - Liu, En Hao

AU - Chen, Jin Yuan

PY - 2007/12/1

Y1 - 2007/12/1

N2 - Branch prediction is a key mechanism to boost the system performance of a superscalar processor. Though the prediction accuracy rate becomes higher and higher, the mispredicitons still lead to significant performance losses in a wide-issue deep-pipelined superscalar. To address the problem, the technique of multipath execution has been proposed previously, which is capable of executing both paths whenever a lower-confidence conditional branch is encountered. However, because the instructions from different paths share a single register update unit (RUU), they are interleaved in the RUU. In consequence, when a conditional branch is resolved and the instructions on the wrong paths are squashed, all the entries in the resulting holes cannot be reused until they are reclaimed at the commit stage. Since the RUU size is crucial to the performance, it is interesting to know how much can we speedup the performance if the squashed RUU entries can be reused as soon as possible. We have proposed a simple mechanism with very limited hardware resources to achieve this goal. Finally the preliminary simulation results are presented.

AB - Branch prediction is a key mechanism to boost the system performance of a superscalar processor. Though the prediction accuracy rate becomes higher and higher, the mispredicitons still lead to significant performance losses in a wide-issue deep-pipelined superscalar. To address the problem, the technique of multipath execution has been proposed previously, which is capable of executing both paths whenever a lower-confidence conditional branch is encountered. However, because the instructions from different paths share a single register update unit (RUU), they are interleaved in the RUU. In consequence, when a conditional branch is resolved and the instructions on the wrong paths are squashed, all the entries in the resulting holes cannot be reused until they are reclaimed at the commit stage. Since the RUU size is crucial to the performance, it is interesting to know how much can we speedup the performance if the squashed RUU entries can be reused as soon as possible. We have proposed a simple mechanism with very limited hardware resources to achieve this goal. Finally the preliminary simulation results are presented.

UR - http://www.scopus.com/inward/record.url?scp=47349087354&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=47349087354&partnerID=8YFLogxK

U2 - 10.1109/PACRIM.2007.4313190

DO - 10.1109/PACRIM.2007.4313190

M3 - Conference contribution

AN - SCOPUS:47349087354

SN - 1424411904

SN - 9781424411900

T3 - IEEE Pacific RIM Conference on Communications, Computers, and Signal Processing - Proceedings

SP - 113

EP - 116

BT - 2007 IEEE Pacific Rim Conference on Communications, Computers and Signal Processing, Conference Proceedings, PACRIM

ER -

Wu C-C, Lai KC, Liu EH, Chen JY. The impact of the register update unit size on multipath execution. In 2007 IEEE Pacific Rim Conference on Communications, Computers and Signal Processing, Conference Proceedings, PACRIM. 2007. p. 113-116. 4313190. (IEEE Pacific RIM Conference on Communications, Computers, and Signal Processing - Proceedings). https://doi.org/10.1109/PACRIM.2007.4313190