Quantifying Overheads in Charm++ and HPX Using Task Bench

Nanmiao Wu; Ioannis Gonidelis; Simeng Liu; Zane Fink; Nikunj Gupta; Karame Mohammadiporshokooh; Patrick Diehl; Hartmut Kaiser; Laxmikant V. Kale

doi:10.1007/978-3-031-31209-0_1

Quantifying Overheads in Charm++ and HPX Using Task Bench

Nanmiao Wu
, Ioannis Gonidelis
, Simeng Liu
, Zane Fink
, Nikunj Gupta
, Karame Mohammadiporshokooh
, Patrick Diehl
, Hartmut Kaiser
, Laxmikant V. Kale

Siebel School of Computing and Data Science

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

Abstract

Asynchronous Many-Task (AMT) runtime systems take advantage of multi-core architectures with light-weight threads, asynchronous executions, and smart scheduling. In this paper, we present the comparison of the AMT systems Charm++ and HPX with the main stream MPI, OpenMP, and MPI+OpenMP libraries using the Task Bench benchmarks. Charm++ is a parallel programming language based on C++, supporting stackless tasks as well as light-weight threads asynchronously along with an adaptive runtime system. HPX is a C++ library for concurrency and parallelism, exposing C++ standards conforming API. First, we analyze the commonalities, differences, and advantageous scenarios of Charm++ and HPX in detail. Further, to investigate the potential overheads introduced by the tasking systems of Charm++ and HPX, we utilize an existing parameterized benchmark, Task Bench, wherein 15 different programming systems were implemented, e.g., MPI, OpenMP, MPI + OpenMP, and extend Task Bench by adding HPX implementations. We quantify the overheads of Charm++, HPX, and the main stream libraries in different scenarios where a single task and multi-task are assigned to each core, respectively. We also investigate each system’s scalability and the ability to hide the communication latency.

Original language	English (US)
Title of host publication	Euro-Par 2022
Subtitle of host publication	Parallel Processing Workshops - Euro-Par 2022 International Workshops, Revised Selected Papers
Editors	Jeremy Singer, Yehia Elkhatib, Dora Blanco Heras, Patrick Diehl, Nick Brown, Aleksandar Ilic
Publisher	Springer
Pages	5-16
Number of pages	12
ISBN (Print)	9783031312083
DOIs	https://doi.org/10.1007/978-3-031-31209-0_1
State	Published - 2023
Event	28th International European Conference on Parallel and Distributed Computing , Euro-Par 2022 - Glasgow, United Kingdom Duration: Aug 22 2022 → Aug 26 2022

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	13835 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	28th International European Conference on Parallel and Distributed Computing , Euro-Par 2022
Country/Territory	United Kingdom
City	Glasgow
Period	8/22/22 → 8/26/22

Keywords

Asynchronous Many-Task (AMT)
Charm++
HPX
Task Bench

ASJC Scopus subject areas

Theoretical Computer Science
General Computer Science

Online availability

10.1007/978-3-031-31209-0_1

Library availability

Discover UIUC Full Text

Cite this

Wu, N., Gonidelis, I., Liu, S., Fink, Z., Gupta, N., Mohammadiporshokooh, K., Diehl, P., Kaiser, H., & Kale, L. V. (2023). Quantifying Overheads in Charm++ and HPX Using Task Bench. In J. Singer, Y. Elkhatib, D. Blanco Heras, P. Diehl, N. Brown, & A. Ilic (Eds.), Euro-Par 2022: Parallel Processing Workshops - Euro-Par 2022 International Workshops, Revised Selected Papers (pp. 5-16). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 13835 LNCS). Springer. https://doi.org/10.1007/978-3-031-31209-0_1

Quantifying Overheads in Charm++ and HPX Using Task Bench. / Wu, Nanmiao; Gonidelis, Ioannis; Liu, Simeng et al.
Euro-Par 2022: Parallel Processing Workshops - Euro-Par 2022 International Workshops, Revised Selected Papers. ed. / Jeremy Singer; Yehia Elkhatib; Dora Blanco Heras; Patrick Diehl; Nick Brown; Aleksandar Ilic. Springer, 2023. p. 5-16 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 13835 LNCS).

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

Wu, N, Gonidelis, I, Liu, S, Fink, Z, Gupta, N, Mohammadiporshokooh, K, Diehl, P, Kaiser, H & Kale, LV 2023, Quantifying Overheads in Charm++ and HPX Using Task Bench. in J Singer, Y Elkhatib, D Blanco Heras, P Diehl, N Brown & A Ilic (eds), Euro-Par 2022: Parallel Processing Workshops - Euro-Par 2022 International Workshops, Revised Selected Papers. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 13835 LNCS, Springer, pp. 5-16, 28th International European Conference on Parallel and Distributed Computing , Euro-Par 2022, Glasgow, United Kingdom, 8/22/22. https://doi.org/10.1007/978-3-031-31209-0_1

Wu N, Gonidelis I, Liu S, Fink Z, Gupta N, Mohammadiporshokooh K et al. Quantifying Overheads in Charm++ and HPX Using Task Bench. In Singer J, Elkhatib Y, Blanco Heras D, Diehl P, Brown N, Ilic A, editors, Euro-Par 2022: Parallel Processing Workshops - Euro-Par 2022 International Workshops, Revised Selected Papers. Springer. 2023. p. 5-16. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-031-31209-0_1

Wu, Nanmiao ; Gonidelis, Ioannis ; Liu, Simeng et al. / Quantifying Overheads in Charm++ and HPX Using Task Bench. Euro-Par 2022: Parallel Processing Workshops - Euro-Par 2022 International Workshops, Revised Selected Papers. editor / Jeremy Singer ; Yehia Elkhatib ; Dora Blanco Heras ; Patrick Diehl ; Nick Brown ; Aleksandar Ilic. Springer, 2023. pp. 5-16 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{1640646dfd5e4d6da07361fd9bb264ba,

title = "Quantifying Overheads in Charm++ and HPX Using Task Bench",

abstract = "Asynchronous Many-Task (AMT) runtime systems take advantage of multi-core architectures with light-weight threads, asynchronous executions, and smart scheduling. In this paper, we present the comparison of the AMT systems Charm++ and HPX with the main stream MPI, OpenMP, and MPI+OpenMP libraries using the Task Bench benchmarks. Charm++ is a parallel programming language based on C++, supporting stackless tasks as well as light-weight threads asynchronously along with an adaptive runtime system. HPX is a C++ library for concurrency and parallelism, exposing C++ standards conforming API. First, we analyze the commonalities, differences, and advantageous scenarios of Charm++ and HPX in detail. Further, to investigate the potential overheads introduced by the tasking systems of Charm++ and HPX, we utilize an existing parameterized benchmark, Task Bench, wherein 15 different programming systems were implemented, e.g., MPI, OpenMP, MPI + OpenMP, and extend Task Bench by adding HPX implementations. We quantify the overheads of Charm++, HPX, and the main stream libraries in different scenarios where a single task and multi-task are assigned to each core, respectively. We also investigate each system{\textquoteright}s scalability and the ability to hide the communication latency.",

keywords = "Asynchronous Many-Task (AMT), Charm++, HPX, Task Bench",

author = "Nanmiao Wu and Ioannis Gonidelis and Simeng Liu and Zane Fink and Nikunj Gupta and Karame Mohammadiporshokooh and Patrick Diehl and Hartmut Kaiser and Kale, \{Laxmikant V.\}",

note = "Publisher Copyright: {\textcopyright} 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.; 28th International European Conference on Parallel and Distributed Computing , Euro-Par 2022 ; Conference date: 22-08-2022 Through 26-08-2022",

year = "2023",

doi = "10.1007/978-3-031-31209-0\_1",

language = "English (US)",

isbn = "9783031312083",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer",

pages = "5--16",

editor = "Jeremy Singer and Yehia Elkhatib and \{Blanco Heras\}, Dora and Patrick Diehl and Nick Brown and Aleksandar Ilic",

booktitle = "Euro-Par 2022",

address = "Germany",

}

TY - GEN

T1 - Quantifying Overheads in Charm++ and HPX Using Task Bench

AU - Wu, Nanmiao

AU - Gonidelis, Ioannis

AU - Liu, Simeng

AU - Fink, Zane

AU - Gupta, Nikunj

AU - Mohammadiporshokooh, Karame

AU - Diehl, Patrick

AU - Kaiser, Hartmut

AU - Kale, Laxmikant V.

PY - 2023

Y1 - 2023

N2 - Asynchronous Many-Task (AMT) runtime systems take advantage of multi-core architectures with light-weight threads, asynchronous executions, and smart scheduling. In this paper, we present the comparison of the AMT systems Charm++ and HPX with the main stream MPI, OpenMP, and MPI+OpenMP libraries using the Task Bench benchmarks. Charm++ is a parallel programming language based on C++, supporting stackless tasks as well as light-weight threads asynchronously along with an adaptive runtime system. HPX is a C++ library for concurrency and parallelism, exposing C++ standards conforming API. First, we analyze the commonalities, differences, and advantageous scenarios of Charm++ and HPX in detail. Further, to investigate the potential overheads introduced by the tasking systems of Charm++ and HPX, we utilize an existing parameterized benchmark, Task Bench, wherein 15 different programming systems were implemented, e.g., MPI, OpenMP, MPI + OpenMP, and extend Task Bench by adding HPX implementations. We quantify the overheads of Charm++, HPX, and the main stream libraries in different scenarios where a single task and multi-task are assigned to each core, respectively. We also investigate each system’s scalability and the ability to hide the communication latency.

AB - Asynchronous Many-Task (AMT) runtime systems take advantage of multi-core architectures with light-weight threads, asynchronous executions, and smart scheduling. In this paper, we present the comparison of the AMT systems Charm++ and HPX with the main stream MPI, OpenMP, and MPI+OpenMP libraries using the Task Bench benchmarks. Charm++ is a parallel programming language based on C++, supporting stackless tasks as well as light-weight threads asynchronously along with an adaptive runtime system. HPX is a C++ library for concurrency and parallelism, exposing C++ standards conforming API. First, we analyze the commonalities, differences, and advantageous scenarios of Charm++ and HPX in detail. Further, to investigate the potential overheads introduced by the tasking systems of Charm++ and HPX, we utilize an existing parameterized benchmark, Task Bench, wherein 15 different programming systems were implemented, e.g., MPI, OpenMP, MPI + OpenMP, and extend Task Bench by adding HPX implementations. We quantify the overheads of Charm++, HPX, and the main stream libraries in different scenarios where a single task and multi-task are assigned to each core, respectively. We also investigate each system’s scalability and the ability to hide the communication latency.

KW - Asynchronous Many-Task (AMT)

KW - Charm++

KW - HPX

KW - Task Bench

UR - https://www.scopus.com/pages/publications/85161451212

UR - https://www.scopus.com/pages/publications/85161451212#tab=citedBy

U2 - 10.1007/978-3-031-31209-0_1

DO - 10.1007/978-3-031-31209-0_1

M3 - Conference contribution

AN - SCOPUS:85161451212

SN - 9783031312083

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 5

EP - 16

BT - Euro-Par 2022

A2 - Singer, Jeremy

A2 - Elkhatib, Yehia

A2 - Blanco Heras, Dora

A2 - Diehl, Patrick

A2 - Brown, Nick

A2 - Ilic, Aleksandar

PB - Springer

T2 - 28th International European Conference on Parallel and Distributed Computing , Euro-Par 2022

Y2 - 22 August 2022 through 26 August 2022

ER -