Context-Aware Memory Profiling for Speculative Parallelism

Changsu Kim; Juhyun Kim; Juwon Kang; Jae W. Lee; Hanjun Kim

doi:10.1109/HiPC.2017.00045

Context-Aware Memory Profiling for Speculative Parallelism

Changsu Kim, Juhyun Kim, Juwon Kang, Jae W. Lee, Hanjun Kim

Department of Electrical and Electronic Engineering

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

1 Citation (Scopus)

Abstract

To expose hidden parallelism from programs with complex dependences, modern compilers employ memory profilers to augment imprecise static analyses. Since dynamic dependence patterns among instructions can vary widely depending on the context, such as function call site stack and loop nest level, context-aware memory profiling is of great value for precise memory profiling. However, recording memory dependences with full context information causes huge overheads in terms of CPU cycles and memory space. Existing profilers mitigate this problem by compromising precision, coverage, or both. This paper proposes a new precise Context-Aware Memory Profiling (CAMP) framework that efficiently traces all the memory dependences with full context information. CAMP statically analyzes a context tree of a program that illustrates all the possible dynamic contexts, and simplifies context management during profiling. For 14 programs from SPEC CINT2000 and CINT2006 benchmark suites, CAMP increases speculative parallelism opportunities by 12.6% on average and by up to 63.0% compared to the baseline context-oblivious, loop-aware memory profiler.

Original language	English
Title of host publication	Proceedings - 24th IEEE International Conference on High Performance Computing, HiPC 2017
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	328-337
Number of pages	10
ISBN (Electronic)	9781538622933
DOIs	https://doi.org/10.1109/HiPC.2017.00045
Publication status	Published - 2018 Feb 7
Event	24th IEEE International Conference on High Performance Computing, HiPC 2017 - Jaipur, India Duration: 2017 Dec 18 → 2017 Dec 21

Publication series

Name	Proceedings - 24th IEEE International Conference on High Performance Computing, HiPC 2017
Volume	2017-December

Other

Other	24th IEEE International Conference on High Performance Computing, HiPC 2017
Country/Territory	India
City	Jaipur
Period	17/12/18 → 17/12/21

Bibliographical note

Funding Information:
We thank the anonymous reviewers for their insightful comments and suggestions. This work is supported by Korea Institute of Science and Technology Information (KISTI) under the grant No.K-16-L01-C03-S03, and the Korean Government (MSIT) under the "Next-Generation Information Computing Development Program" (NRF-2015M3C4A7065646), the "Basic Science Research Program" (NRF-2017R1C1B3009332), the "ICT ConsiUence Creative Program" (IITP-2017-R0346-16-1007), and the "PF Class Heterogeneous High Performance Computer Development" (NRF-2016M3C4A7952587).

Publisher Copyright:
© 2017 IEEE.

All Science Journal Classification (ASJC) codes

Computer Networks and Communications
Hardware and Architecture
Modelling and Simulation

Access to Document

10.1109/HiPC.2017.00045

Cite this

Kim, C., Kim, J., Kang, J., Lee, J. W., & Kim, H. (2018). Context-Aware Memory Profiling for Speculative Parallelism. In Proceedings - 24th IEEE International Conference on High Performance Computing, HiPC 2017 (pp. 328-337). (Proceedings - 24th IEEE International Conference on High Performance Computing, HiPC 2017; Vol. 2017-December). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/HiPC.2017.00045

@inproceedings{2984a06e5a8240d7a25e07b54f5f7213,

title = "Context-Aware Memory Profiling for Speculative Parallelism",

abstract = "To expose hidden parallelism from programs with complex dependences, modern compilers employ memory profilers to augment imprecise static analyses. Since dynamic dependence patterns among instructions can vary widely depending on the context, such as function call site stack and loop nest level, context-aware memory profiling is of great value for precise memory profiling. However, recording memory dependences with full context information causes huge overheads in terms of CPU cycles and memory space. Existing profilers mitigate this problem by compromising precision, coverage, or both. This paper proposes a new precise Context-Aware Memory Profiling (CAMP) framework that efficiently traces all the memory dependences with full context information. CAMP statically analyzes a context tree of a program that illustrates all the possible dynamic contexts, and simplifies context management during profiling. For 14 programs from SPEC CINT2000 and CINT2006 benchmark suites, CAMP increases speculative parallelism opportunities by 12.6% on average and by up to 63.0% compared to the baseline context-oblivious, loop-aware memory profiler.",

author = "Changsu Kim and Juhyun Kim and Juwon Kang and Lee, {Jae W.} and Hanjun Kim",

note = "Funding Information: We thank the anonymous reviewers for their insightful comments and suggestions. This work is supported by Korea Institute of Science and Technology Information (KISTI) under the grant No.K-16-L01-C03-S03, and the Korean Government (MSIT) under the {"}Next-Generation Information Computing Development Program{"} (NRF-2015M3C4A7065646), the {"}Basic Science Research Program{"} (NRF-2017R1C1B3009332), the {"}ICT ConsiUence Creative Program{"} (IITP-2017-R0346-16-1007), and the {"}PF Class Heterogeneous High Performance Computer Development{"} (NRF-2016M3C4A7952587). Publisher Copyright: {\textcopyright} 2017 IEEE.; 24th IEEE International Conference on High Performance Computing, HiPC 2017 ; Conference date: 18-12-2017 Through 21-12-2017",

year = "2018",

month = feb,

day = "7",

doi = "10.1109/HiPC.2017.00045",

language = "English",

series = "Proceedings - 24th IEEE International Conference on High Performance Computing, HiPC 2017",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

pages = "328--337",

booktitle = "Proceedings - 24th IEEE International Conference on High Performance Computing, HiPC 2017",

address = "United States",

}

Kim, C, Kim, J, Kang, J, Lee, JW & Kim, H 2018, Context-Aware Memory Profiling for Speculative Parallelism. in Proceedings - 24th IEEE International Conference on High Performance Computing, HiPC 2017. Proceedings - 24th IEEE International Conference on High Performance Computing, HiPC 2017, vol. 2017-December, Institute of Electrical and Electronics Engineers Inc., pp. 328-337, 24th IEEE International Conference on High Performance Computing, HiPC 2017, Jaipur, India, 17/12/18. https://doi.org/10.1109/HiPC.2017.00045

Context-Aware Memory Profiling for Speculative Parallelism. / Kim, Changsu; Kim, Juhyun; Kang, Juwon et al.
Proceedings - 24th IEEE International Conference on High Performance Computing, HiPC 2017. Institute of Electrical and Electronics Engineers Inc., 2018. p. 328-337 (Proceedings - 24th IEEE International Conference on High Performance Computing, HiPC 2017; Vol. 2017-December).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

TY - GEN

T1 - Context-Aware Memory Profiling for Speculative Parallelism

AU - Kim, Changsu

AU - Kim, Juhyun

AU - Kang, Juwon

AU - Lee, Jae W.

AU - Kim, Hanjun

N1 - Funding Information: We thank the anonymous reviewers for their insightful comments and suggestions. This work is supported by Korea Institute of Science and Technology Information (KISTI) under the grant No.K-16-L01-C03-S03, and the Korean Government (MSIT) under the "Next-Generation Information Computing Development Program" (NRF-2015M3C4A7065646), the "Basic Science Research Program" (NRF-2017R1C1B3009332), the "ICT ConsiUence Creative Program" (IITP-2017-R0346-16-1007), and the "PF Class Heterogeneous High Performance Computer Development" (NRF-2016M3C4A7952587). Publisher Copyright: © 2017 IEEE.

PY - 2018/2/7

Y1 - 2018/2/7

N2 - To expose hidden parallelism from programs with complex dependences, modern compilers employ memory profilers to augment imprecise static analyses. Since dynamic dependence patterns among instructions can vary widely depending on the context, such as function call site stack and loop nest level, context-aware memory profiling is of great value for precise memory profiling. However, recording memory dependences with full context information causes huge overheads in terms of CPU cycles and memory space. Existing profilers mitigate this problem by compromising precision, coverage, or both. This paper proposes a new precise Context-Aware Memory Profiling (CAMP) framework that efficiently traces all the memory dependences with full context information. CAMP statically analyzes a context tree of a program that illustrates all the possible dynamic contexts, and simplifies context management during profiling. For 14 programs from SPEC CINT2000 and CINT2006 benchmark suites, CAMP increases speculative parallelism opportunities by 12.6% on average and by up to 63.0% compared to the baseline context-oblivious, loop-aware memory profiler.

AB - To expose hidden parallelism from programs with complex dependences, modern compilers employ memory profilers to augment imprecise static analyses. Since dynamic dependence patterns among instructions can vary widely depending on the context, such as function call site stack and loop nest level, context-aware memory profiling is of great value for precise memory profiling. However, recording memory dependences with full context information causes huge overheads in terms of CPU cycles and memory space. Existing profilers mitigate this problem by compromising precision, coverage, or both. This paper proposes a new precise Context-Aware Memory Profiling (CAMP) framework that efficiently traces all the memory dependences with full context information. CAMP statically analyzes a context tree of a program that illustrates all the possible dynamic contexts, and simplifies context management during profiling. For 14 programs from SPEC CINT2000 and CINT2006 benchmark suites, CAMP increases speculative parallelism opportunities by 12.6% on average and by up to 63.0% compared to the baseline context-oblivious, loop-aware memory profiler.

UR - http://www.scopus.com/inward/record.url?scp=85050350296&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85050350296&partnerID=8YFLogxK

U2 - 10.1109/HiPC.2017.00045

DO - 10.1109/HiPC.2017.00045

M3 - Conference contribution

AN - SCOPUS:85050350296

T3 - Proceedings - 24th IEEE International Conference on High Performance Computing, HiPC 2017

SP - 328

EP - 337

BT - Proceedings - 24th IEEE International Conference on High Performance Computing, HiPC 2017

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 24th IEEE International Conference on High Performance Computing, HiPC 2017

Y2 - 18 December 2017 through 21 December 2017

ER -

Kim C, Kim J, Kang J, Lee JW, Kim H. Context-Aware Memory Profiling for Speculative Parallelism. In Proceedings - 24th IEEE International Conference on High Performance Computing, HiPC 2017. Institute of Electrical and Electronics Engineers Inc. 2018. p. 328-337. (Proceedings - 24th IEEE International Conference on High Performance Computing, HiPC 2017). doi: 10.1109/HiPC.2017.00045

Context-Aware Memory Profiling for Speculative Parallelism

Abstract

Publication series

Other

Bibliographical note

All Science Journal Classification (ASJC) codes

Access to Document

Other files and links

Fingerprint

Cite this