A taxonomy of task-based parallel programming technologies for high-performance computing

Peter Thoman; Kiril Dichev; Thomas Heller; Roman  Iakymchuk; Xavier  Aguilar; Khalid  Hasanov; Philipp  Gschwandtner; Pierre  Lemarinier; Stefano  Markidis; Herbert Jordan; Thomas Fahringer; Kostas  Katrinis; Erwin  Laure; Dimitrios Nikolopoulos

doi:10.1007/s11227-018-2238-4

A taxonomy of task-based parallel programming technologies for high-performance computing

Peter Thoman
, Kiril Dichev
, Thomas Heller
, Roman Iakymchuk
, Xavier Aguilar
, Khalid Hasanov
, Philipp Gschwandtner
, Pierre Lemarinier
, Stefano Markidis
, Herbert Jordan
, Thomas Fahringer
, Kostas Katrinis
, Erwin Laure
, Dimitrios Nikolopoulos

School of Electronics, Electrical Engineering and Computer Science

Research output: Contribution to journal › Article › peer-review

122 Citations (Scopus)

528 Downloads (Pure)

Abstract

Task-based programming models for shared memory - such as Cilk Plus and OpenMP 3 - are well established and documented. However, with the increase in parallel, many-core, and heterogeneous systems, a number of research-driven projects have developed more diversified task-based support, employing various programming and runtime features. Unfortunately, despite the fact that dozens of different task-based systems exist today and are actively used for parallel and high-performance computing (HPC), no comprehensive overview or classification of task-based technologies for HPC exists. In this paper, we provide an initial task-focused taxonomy for HPC technologies, which covers both programming interfaces and runtime mechanisms. We demonstrate the usefulness of our taxonomy by classifying state-of-the-art task-based environments in use today.

Original language	English
Pages (from-to)	1-13
Journal	The Journal of Supercomputing
DOIs	https://doi.org/10.1007/s11227-018-2238-4
Publication status	Published - 12 Jan 2018

Keywords

High-performance computing
Task-based parallelism
Taxonomy
API
Runtime system
Scheduler
Monitoring framework
Fault tolerance

Access to Document

10.1007/s11227-018-2238-4Licence: CC BY

A taxonomy of task-based parallel programming technologies for high-performance computing
Copyright 2018 the authors. This is an open access article published under a Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution and reproduction in any medium, provided the author and source are cited.
Final published version, 573 KBLicence: CC BY

Cite this

Thoman, P., Dichev, K., Heller, T., Iakymchuk, R., Aguilar, X., Hasanov, K., Gschwandtner, P., Lemarinier, P., Markidis, S., Jordan, H., Fahringer, T., Katrinis, K., Laure, E., & Nikolopoulos, D. (2018). A taxonomy of task-based parallel programming technologies for high-performance computing. The Journal of Supercomputing, 1-13. https://doi.org/10.1007/s11227-018-2238-4

@article{0941a76394344a648391846bbff39b88,

title = "A taxonomy of task-based parallel programming technologies for high-performance computing",

abstract = "Task-based programming models for shared memory - such as Cilk Plus and OpenMP 3 - are well established and documented. However, with the increase in parallel, many-core, and heterogeneous systems, a number of research-driven projects have developed more diversified task-based support, employing various programming and runtime features. Unfortunately, despite the fact that dozens of different task-based systems exist today and are actively used for parallel and high-performance computing (HPC), no comprehensive overview or classification of task-based technologies for HPC exists. In this paper, we provide an initial task-focused taxonomy for HPC technologies, which covers both programming interfaces and runtime mechanisms. We demonstrate the usefulness of our taxonomy by classifying state-of-the-art task-based environments in use today.",

keywords = "High-performance computing, Task-based parallelism, Taxonomy, API, Runtime system, Scheduler, Monitoring framework, Fault tolerance",

author = "Peter Thoman and Kiril Dichev and Thomas Heller and Roman Iakymchuk and Xavier Aguilar and Khalid Hasanov and Philipp Gschwandtner and Pierre Lemarinier and Stefano Markidis and Herbert Jordan and Thomas Fahringer and Kostas Katrinis and Erwin Laure and Dimitrios Nikolopoulos",

year = "2018",

month = jan,

day = "12",

doi = "10.1007/s11227-018-2238-4",

language = "English",

pages = "1--13",

journal = "The Journal of Supercomputing",

issn = "0920-8542",

publisher = "Springer Netherlands",

}

TY - JOUR

T1 - A taxonomy of task-based parallel programming technologies for high-performance computing

AU - Thoman, Peter

AU - Dichev, Kiril

AU - Heller, Thomas

AU - Iakymchuk, Roman

AU - Aguilar, Xavier

AU - Hasanov, Khalid

AU - Gschwandtner, Philipp

AU - Lemarinier, Pierre

AU - Markidis, Stefano

AU - Jordan, Herbert

AU - Fahringer, Thomas

AU - Katrinis, Kostas

AU - Laure, Erwin

AU - Nikolopoulos, Dimitrios

PY - 2018/1/12

Y1 - 2018/1/12

N2 - Task-based programming models for shared memory - such as Cilk Plus and OpenMP 3 - are well established and documented. However, with the increase in parallel, many-core, and heterogeneous systems, a number of research-driven projects have developed more diversified task-based support, employing various programming and runtime features. Unfortunately, despite the fact that dozens of different task-based systems exist today and are actively used for parallel and high-performance computing (HPC), no comprehensive overview or classification of task-based technologies for HPC exists. In this paper, we provide an initial task-focused taxonomy for HPC technologies, which covers both programming interfaces and runtime mechanisms. We demonstrate the usefulness of our taxonomy by classifying state-of-the-art task-based environments in use today.

AB - Task-based programming models for shared memory - such as Cilk Plus and OpenMP 3 - are well established and documented. However, with the increase in parallel, many-core, and heterogeneous systems, a number of research-driven projects have developed more diversified task-based support, employing various programming and runtime features. Unfortunately, despite the fact that dozens of different task-based systems exist today and are actively used for parallel and high-performance computing (HPC), no comprehensive overview or classification of task-based technologies for HPC exists. In this paper, we provide an initial task-focused taxonomy for HPC technologies, which covers both programming interfaces and runtime mechanisms. We demonstrate the usefulness of our taxonomy by classifying state-of-the-art task-based environments in use today.

KW - High-performance computing

KW - Task-based parallelism

KW - Taxonomy

KW - API

KW - Runtime system

KW - Scheduler

KW - Monitoring framework

KW - Fault tolerance

U2 - 10.1007/s11227-018-2238-4

DO - 10.1007/s11227-018-2238-4

M3 - Article

SN - 0920-8542

SP - 1

EP - 13

JO - The Journal of Supercomputing

JF - The Journal of Supercomputing

ER -

A taxonomy of task-based parallel programming technologies for high-performance computing

Abstract

Keywords

Access to Document

Fingerprint

Cite this