Efficient pointer management of stack data for software managed multicores

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

Abstract

Scratchpad-memory (SPM) based memory hierarchy is a promising alternative to cache-based memory hierarchies, due to the difficulty in scaling caches to processors with high core count. However, explicit data management in software is required on SPM-based memory hierarchies. This paper focuses on optimizing the stack data management on SPM-based multicore processors, as memory accesses to call stack present in most applications. While previous works have developed techniques to enable correct stack pointer management, they have not optimized it. As a result, existing techniques still incur high overhead. This paper proposes an automated compiler-based scheme for efficient pointer management. Our experiments on MiBench benchmarks demonstrate that our scheme almost completely eliminates pointer management overhead. As a result, as compared to the state-of-the-art approach, our approach reduces the average execution time of the benchmarks by 52%. Furthermore, with our approach, the performance of stack management on SPM becomes better than hardware caches on average even with conservative estimates.

Original languageEnglish (US)
Title of host publication2016 IEEE 27th International Conference on Application-Specific Systems, Architectures and Processors, ASAP 2016
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages67-74
Number of pages8
Volume2016-November
ISBN (Electronic)9781509015030
DOIs
StatePublished - Nov 28 2016
Event27th IEEE International Conference on Application-Specific Systems, Architectures and Processors, ASAP 2016 - London, United Kingdom
Duration: Jul 6 2016Jul 8 2016

Other

Other27th IEEE International Conference on Application-Specific Systems, Architectures and Processors, ASAP 2016
CountryUnited Kingdom
CityLondon
Period7/6/167/8/16

Fingerprint

Data storage equipment
Information management
Computer hardware
Experiments

ASJC Scopus subject areas

  • Hardware and Architecture
  • Computer Networks and Communications

Cite this

Cai, J., & Shrivastava, A. (2016). Efficient pointer management of stack data for software managed multicores. In 2016 IEEE 27th International Conference on Application-Specific Systems, Architectures and Processors, ASAP 2016 (Vol. 2016-November, pp. 67-74). [7760774] Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ASAP.2016.7760774

Efficient pointer management of stack data for software managed multicores. / Cai, Jian; Shrivastava, Aviral.

2016 IEEE 27th International Conference on Application-Specific Systems, Architectures and Processors, ASAP 2016. Vol. 2016-November Institute of Electrical and Electronics Engineers Inc., 2016. p. 67-74 7760774.

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

Cai, J & Shrivastava, A 2016, Efficient pointer management of stack data for software managed multicores. in 2016 IEEE 27th International Conference on Application-Specific Systems, Architectures and Processors, ASAP 2016. vol. 2016-November, 7760774, Institute of Electrical and Electronics Engineers Inc., pp. 67-74, 27th IEEE International Conference on Application-Specific Systems, Architectures and Processors, ASAP 2016, London, United Kingdom, 7/6/16. https://doi.org/10.1109/ASAP.2016.7760774
Cai J, Shrivastava A. Efficient pointer management of stack data for software managed multicores. In 2016 IEEE 27th International Conference on Application-Specific Systems, Architectures and Processors, ASAP 2016. Vol. 2016-November. Institute of Electrical and Electronics Engineers Inc. 2016. p. 67-74. 7760774 https://doi.org/10.1109/ASAP.2016.7760774
Cai, Jian ; Shrivastava, Aviral. / Efficient pointer management of stack data for software managed multicores. 2016 IEEE 27th International Conference on Application-Specific Systems, Architectures and Processors, ASAP 2016. Vol. 2016-November Institute of Electrical and Electronics Engineers Inc., 2016. pp. 67-74
@inproceedings{f118416339f545fdbe867c4ff5f6b554,
title = "Efficient pointer management of stack data for software managed multicores",
abstract = "Scratchpad-memory (SPM) based memory hierarchy is a promising alternative to cache-based memory hierarchies, due to the difficulty in scaling caches to processors with high core count. However, explicit data management in software is required on SPM-based memory hierarchies. This paper focuses on optimizing the stack data management on SPM-based multicore processors, as memory accesses to call stack present in most applications. While previous works have developed techniques to enable correct stack pointer management, they have not optimized it. As a result, existing techniques still incur high overhead. This paper proposes an automated compiler-based scheme for efficient pointer management. Our experiments on MiBench benchmarks demonstrate that our scheme almost completely eliminates pointer management overhead. As a result, as compared to the state-of-the-art approach, our approach reduces the average execution time of the benchmarks by 52{\%}. Furthermore, with our approach, the performance of stack management on SPM becomes better than hardware caches on average even with conservative estimates.",
author = "Jian Cai and Aviral Shrivastava",
year = "2016",
month = "11",
day = "28",
doi = "10.1109/ASAP.2016.7760774",
language = "English (US)",
volume = "2016-November",
pages = "67--74",
booktitle = "2016 IEEE 27th International Conference on Application-Specific Systems, Architectures and Processors, ASAP 2016",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
address = "United States",

}

TY - GEN

T1 - Efficient pointer management of stack data for software managed multicores

AU - Cai, Jian

AU - Shrivastava, Aviral

PY - 2016/11/28

Y1 - 2016/11/28

N2 - Scratchpad-memory (SPM) based memory hierarchy is a promising alternative to cache-based memory hierarchies, due to the difficulty in scaling caches to processors with high core count. However, explicit data management in software is required on SPM-based memory hierarchies. This paper focuses on optimizing the stack data management on SPM-based multicore processors, as memory accesses to call stack present in most applications. While previous works have developed techniques to enable correct stack pointer management, they have not optimized it. As a result, existing techniques still incur high overhead. This paper proposes an automated compiler-based scheme for efficient pointer management. Our experiments on MiBench benchmarks demonstrate that our scheme almost completely eliminates pointer management overhead. As a result, as compared to the state-of-the-art approach, our approach reduces the average execution time of the benchmarks by 52%. Furthermore, with our approach, the performance of stack management on SPM becomes better than hardware caches on average even with conservative estimates.

AB - Scratchpad-memory (SPM) based memory hierarchy is a promising alternative to cache-based memory hierarchies, due to the difficulty in scaling caches to processors with high core count. However, explicit data management in software is required on SPM-based memory hierarchies. This paper focuses on optimizing the stack data management on SPM-based multicore processors, as memory accesses to call stack present in most applications. While previous works have developed techniques to enable correct stack pointer management, they have not optimized it. As a result, existing techniques still incur high overhead. This paper proposes an automated compiler-based scheme for efficient pointer management. Our experiments on MiBench benchmarks demonstrate that our scheme almost completely eliminates pointer management overhead. As a result, as compared to the state-of-the-art approach, our approach reduces the average execution time of the benchmarks by 52%. Furthermore, with our approach, the performance of stack management on SPM becomes better than hardware caches on average even with conservative estimates.

UR - http://www.scopus.com/inward/record.url?scp=85006922642&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85006922642&partnerID=8YFLogxK

U2 - 10.1109/ASAP.2016.7760774

DO - 10.1109/ASAP.2016.7760774

M3 - Conference contribution

AN - SCOPUS:85006922642

VL - 2016-November

SP - 67

EP - 74

BT - 2016 IEEE 27th International Conference on Application-Specific Systems, Architectures and Processors, ASAP 2016

PB - Institute of Electrical and Electronics Engineers Inc.

ER -