Artificial bee colony-inspired run-time task management for many-core systems

Research output: Contribution to conferencePaper

Standard

Artificial bee colony-inspired run-time task management for many-core systems. / Abuassal, Ali Mohamed Ahmed; Tempesti, Gianluca; Trefzer, Martin Albrecht.

2018. 1084-1091 Paper presented at 2018 SYMPOSIUM SERIES ON COMPUTATIONAL INTELLIGENCE, Bengaluru, India.

Research output: Contribution to conferencePaper

Harvard

Abuassal, AMA, Tempesti, G & Trefzer, MA 2018, 'Artificial bee colony-inspired run-time task management for many-core systems' Paper presented at 2018 SYMPOSIUM SERIES ON COMPUTATIONAL INTELLIGENCE, Bengaluru, India, 18/11/18 - 21/11/18, pp. 1084-1091.

APA

Abuassal, A. M. A., Tempesti, G., & Trefzer, M. A. (2018). Artificial bee colony-inspired run-time task management for many-core systems. 1084-1091. Paper presented at 2018 SYMPOSIUM SERIES ON COMPUTATIONAL INTELLIGENCE, Bengaluru, India.

Vancouver

Abuassal AMA, Tempesti G, Trefzer MA. Artificial bee colony-inspired run-time task management for many-core systems. 2018. Paper presented at 2018 SYMPOSIUM SERIES ON COMPUTATIONAL INTELLIGENCE, Bengaluru, India.

Author

Abuassal, Ali Mohamed Ahmed ; Tempesti, Gianluca ; Trefzer, Martin Albrecht. / Artificial bee colony-inspired run-time task management for many-core systems. Paper presented at 2018 SYMPOSIUM SERIES ON COMPUTATIONAL INTELLIGENCE, Bengaluru, India.8 p.

Bibtex - Download

@conference{83f1aad73e264609b0df03af36d0d330,
title = "Artificial bee colony-inspired run-time task management for many-core systems",
abstract = "Efficient resource and application managementis one of the most complex and challenging tasks in highperformance computing. Large-scale computing systemsthat contain hundreds, thousands or even millions of coresdemand solutions that can operate in a distributed, robust,and scalable fashion. However, while hardware parallelismis relatively straight forward to achieve, this is not generallythe case for software. This leads to under-utilization of thehardware parallelism as well as imbalanced load distributioncausing inefficiency and hotspots. In response to this challenge,this paper introduces a novel distributed and decentralizedrun-time management algorithm. The proposed method isguided by an optimization model inspired by artificial beecolonies (ABC). While ABC have proven useful for optimizinglarge sets of numerical test functions, this is the first time theyare applied in the context of many-core system management.The initial result shows that, the ABC model is promising incontext of run-time management for many-core systems. It isalso anticipated that the algorithms bio-inspired foundationswill inherently enable scalability, reliability, and adaptation.We are showing initial experiments, where the initial resultsindicate the capability of our model to improve the thermaldistribution across the system.",
keywords = "Many-core system, Bio-inspired Hardware, bee colony, run-time management",
author = "Abuassal, {Ali Mohamed Ahmed} and Gianluca Tempesti and Trefzer, {Martin Albrecht}",
note = "{\circledC} IEEE, 2018. This is an author-produced version of the published paper. Uploaded in accordance with the publisher’s self-archiving policy. Further copying may not be permitted; contact the publisher for details. ; 2018 SYMPOSIUM SERIES ON COMPUTATIONAL INTELLIGENCE, IEEE-SSCI 2018 ; Conference date: 18-11-2018 Through 21-11-2018",
year = "2018",
month = "11",
language = "English",
pages = "1084--1091",
url = "http://ieee-ssci2018.org/",

}

RIS (suitable for import to EndNote) - Download

TY - CONF

T1 - Artificial bee colony-inspired run-time task management for many-core systems

AU - Abuassal, Ali Mohamed Ahmed

AU - Tempesti, Gianluca

AU - Trefzer, Martin Albrecht

N1 - © IEEE, 2018. This is an author-produced version of the published paper. Uploaded in accordance with the publisher’s self-archiving policy. Further copying may not be permitted; contact the publisher for details.

PY - 2018/11

Y1 - 2018/11

N2 - Efficient resource and application managementis one of the most complex and challenging tasks in highperformance computing. Large-scale computing systemsthat contain hundreds, thousands or even millions of coresdemand solutions that can operate in a distributed, robust,and scalable fashion. However, while hardware parallelismis relatively straight forward to achieve, this is not generallythe case for software. This leads to under-utilization of thehardware parallelism as well as imbalanced load distributioncausing inefficiency and hotspots. In response to this challenge,this paper introduces a novel distributed and decentralizedrun-time management algorithm. The proposed method isguided by an optimization model inspired by artificial beecolonies (ABC). While ABC have proven useful for optimizinglarge sets of numerical test functions, this is the first time theyare applied in the context of many-core system management.The initial result shows that, the ABC model is promising incontext of run-time management for many-core systems. It isalso anticipated that the algorithms bio-inspired foundationswill inherently enable scalability, reliability, and adaptation.We are showing initial experiments, where the initial resultsindicate the capability of our model to improve the thermaldistribution across the system.

AB - Efficient resource and application managementis one of the most complex and challenging tasks in highperformance computing. Large-scale computing systemsthat contain hundreds, thousands or even millions of coresdemand solutions that can operate in a distributed, robust,and scalable fashion. However, while hardware parallelismis relatively straight forward to achieve, this is not generallythe case for software. This leads to under-utilization of thehardware parallelism as well as imbalanced load distributioncausing inefficiency and hotspots. In response to this challenge,this paper introduces a novel distributed and decentralizedrun-time management algorithm. The proposed method isguided by an optimization model inspired by artificial beecolonies (ABC). While ABC have proven useful for optimizinglarge sets of numerical test functions, this is the first time theyare applied in the context of many-core system management.The initial result shows that, the ABC model is promising incontext of run-time management for many-core systems. It isalso anticipated that the algorithms bio-inspired foundationswill inherently enable scalability, reliability, and adaptation.We are showing initial experiments, where the initial resultsindicate the capability of our model to improve the thermaldistribution across the system.

KW - Many-core system

KW - Bio-inspired Hardware

KW - bee colony

KW - run-time management

M3 - Paper

SP - 1084

EP - 1091

ER -