Reducing code management overhead in software-managed multicores

Jian Cai, Yooseong Kim, Youngbin Kim, Aviral Shrivastava, Kyoungwoo Lee

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

Abstract

Software-managed architectures, which use scratch-pad memories (SPMs), are a promising alternative to cached-based architectures for multicores. SPMs provide scalability but require explicit management. For example, to use an instruction SPM, explicit management code needs to be inserted around every call site to load functions to the SPM. such management code would check the state of the SPM and perform loading operations if necessary, which can cause considerable overhead at runtime. In this paper, we propose a compiler-based approach to reduce this overhead by identifying management code that can be removed or simplified. Our experiments with various benchmarks show that our approach reduces the execution time by 14% on average. In addition, compared to hardware caching, using our approach on an SPM-based architecture can reduce the execution times of the benchmarks by up to 15%.

Original languageEnglish
Title of host publicationProceedings of the 2017 Design, Automation and Test in Europe, DATE 2017
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages1241-1244
Number of pages4
ISBN (Electronic)9783981537093
DOIs
Publication statusPublished - 2017 May 11
Event20th Design, Automation and Test in Europe, DATE 2017 - Swisstech, Lausanne, Switzerland
Duration: 2017 Mar 272017 Mar 31

Publication series

NameProceedings of the 2017 Design, Automation and Test in Europe, DATE 2017

Other

Other20th Design, Automation and Test in Europe, DATE 2017
CountrySwitzerland
CitySwisstech, Lausanne
Period17/3/2717/3/31

Fingerprint

Data storage equipment
Software architecture
Computer hardware
Scalability
Experiments

All Science Journal Classification (ASJC) codes

  • Computer Networks and Communications
  • Hardware and Architecture
  • Safety, Risk, Reliability and Quality

Cite this

Cai, J., Kim, Y., Kim, Y., Shrivastava, A., & Lee, K. (2017). Reducing code management overhead in software-managed multicores. In Proceedings of the 2017 Design, Automation and Test in Europe, DATE 2017 (pp. 1241-1244). [7927179] (Proceedings of the 2017 Design, Automation and Test in Europe, DATE 2017). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.23919/DATE.2017.7927179
Cai, Jian ; Kim, Yooseong ; Kim, Youngbin ; Shrivastava, Aviral ; Lee, Kyoungwoo. / Reducing code management overhead in software-managed multicores. Proceedings of the 2017 Design, Automation and Test in Europe, DATE 2017. Institute of Electrical and Electronics Engineers Inc., 2017. pp. 1241-1244 (Proceedings of the 2017 Design, Automation and Test in Europe, DATE 2017).
@inproceedings{6096f1b3616f4f57a15b2476635db5ec,
title = "Reducing code management overhead in software-managed multicores",
abstract = "Software-managed architectures, which use scratch-pad memories (SPMs), are a promising alternative to cached-based architectures for multicores. SPMs provide scalability but require explicit management. For example, to use an instruction SPM, explicit management code needs to be inserted around every call site to load functions to the SPM. such management code would check the state of the SPM and perform loading operations if necessary, which can cause considerable overhead at runtime. In this paper, we propose a compiler-based approach to reduce this overhead by identifying management code that can be removed or simplified. Our experiments with various benchmarks show that our approach reduces the execution time by 14{\%} on average. In addition, compared to hardware caching, using our approach on an SPM-based architecture can reduce the execution times of the benchmarks by up to 15{\%}.",
author = "Jian Cai and Yooseong Kim and Youngbin Kim and Aviral Shrivastava and Kyoungwoo Lee",
year = "2017",
month = "5",
day = "11",
doi = "10.23919/DATE.2017.7927179",
language = "English",
series = "Proceedings of the 2017 Design, Automation and Test in Europe, DATE 2017",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
pages = "1241--1244",
booktitle = "Proceedings of the 2017 Design, Automation and Test in Europe, DATE 2017",
address = "United States",

}

Cai, J, Kim, Y, Kim, Y, Shrivastava, A & Lee, K 2017, Reducing code management overhead in software-managed multicores. in Proceedings of the 2017 Design, Automation and Test in Europe, DATE 2017., 7927179, Proceedings of the 2017 Design, Automation and Test in Europe, DATE 2017, Institute of Electrical and Electronics Engineers Inc., pp. 1241-1244, 20th Design, Automation and Test in Europe, DATE 2017, Swisstech, Lausanne, Switzerland, 17/3/27. https://doi.org/10.23919/DATE.2017.7927179

Reducing code management overhead in software-managed multicores. / Cai, Jian; Kim, Yooseong; Kim, Youngbin; Shrivastava, Aviral; Lee, Kyoungwoo.

Proceedings of the 2017 Design, Automation and Test in Europe, DATE 2017. Institute of Electrical and Electronics Engineers Inc., 2017. p. 1241-1244 7927179 (Proceedings of the 2017 Design, Automation and Test in Europe, DATE 2017).

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

TY - GEN

T1 - Reducing code management overhead in software-managed multicores

AU - Cai, Jian

AU - Kim, Yooseong

AU - Kim, Youngbin

AU - Shrivastava, Aviral

AU - Lee, Kyoungwoo

PY - 2017/5/11

Y1 - 2017/5/11

N2 - Software-managed architectures, which use scratch-pad memories (SPMs), are a promising alternative to cached-based architectures for multicores. SPMs provide scalability but require explicit management. For example, to use an instruction SPM, explicit management code needs to be inserted around every call site to load functions to the SPM. such management code would check the state of the SPM and perform loading operations if necessary, which can cause considerable overhead at runtime. In this paper, we propose a compiler-based approach to reduce this overhead by identifying management code that can be removed or simplified. Our experiments with various benchmarks show that our approach reduces the execution time by 14% on average. In addition, compared to hardware caching, using our approach on an SPM-based architecture can reduce the execution times of the benchmarks by up to 15%.

AB - Software-managed architectures, which use scratch-pad memories (SPMs), are a promising alternative to cached-based architectures for multicores. SPMs provide scalability but require explicit management. For example, to use an instruction SPM, explicit management code needs to be inserted around every call site to load functions to the SPM. such management code would check the state of the SPM and perform loading operations if necessary, which can cause considerable overhead at runtime. In this paper, we propose a compiler-based approach to reduce this overhead by identifying management code that can be removed or simplified. Our experiments with various benchmarks show that our approach reduces the execution time by 14% on average. In addition, compared to hardware caching, using our approach on an SPM-based architecture can reduce the execution times of the benchmarks by up to 15%.

UR - http://www.scopus.com/inward/record.url?scp=85020237141&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85020237141&partnerID=8YFLogxK

U2 - 10.23919/DATE.2017.7927179

DO - 10.23919/DATE.2017.7927179

M3 - Conference contribution

T3 - Proceedings of the 2017 Design, Automation and Test in Europe, DATE 2017

SP - 1241

EP - 1244

BT - Proceedings of the 2017 Design, Automation and Test in Europe, DATE 2017

PB - Institute of Electrical and Electronics Engineers Inc.

ER -

Cai J, Kim Y, Kim Y, Shrivastava A, Lee K. Reducing code management overhead in software-managed multicores. In Proceedings of the 2017 Design, Automation and Test in Europe, DATE 2017. Institute of Electrical and Electronics Engineers Inc. 2017. p. 1241-1244. 7927179. (Proceedings of the 2017 Design, Automation and Test in Europe, DATE 2017). https://doi.org/10.23919/DATE.2017.7927179