Adaptive scheduling of collocated applications using a task-based runtime system

Jiri Dokulil; Siegfried Benkner

doi:10.1109/CAHPC.2018.8645869

Adaptive scheduling of collocated applications using a task-based runtime system

Jiri Dokulil
, Siegfried Benkner

Research Group Scientific Computing

Publications: Contribution to book › Contribution to proceedings › Peer Reviewed

Abstract

Task-based runtime systems are considered as one of the options for dealing with the challenges of upcoming parallel architectures. The greater flexibility of these runtime systems can also be used to dynamically adjust the resources allocated to the applications, adapting to the current load of the system and the progress of the applications. In our work, we have extended our implementation of the Open Community Runtime to support dynamic adjustment of execution threads. The runtimes communicate with an agent process, which collects performance data, computes thread allocation, and instructs the runtimes to make the required adjustments. We have tested our solution under different scenarios, focusing on producer-consumer applications, where the dynamic resource management was used to keep the applications in sync, improving the overall performance in some cases.

Original language	English
Title of host publication	2018 30th International Symposium on Computer Architecture and High Performance Computing
Publisher	IEEE
Pages	41-48
Number of pages	8
ISBN (Electronic)	978-1-5386-7769-8
DOIs	https://doi.org/10.1109/CAHPC.2018.8645869
Publication status	Published - 2 Jul 2018
Event	30th International Symposium on Computer Architecture and High Performance Computing - École Normale Supérieure, Lyon, France Duration: 24 Sept 2018 → 27 Sept 2018 https://graal.ens-lyon.fr/sbac-pad/

Conference

Conference	30th International Symposium on Computer Architecture and High Performance Computing
Abbreviated title	SBAC-PAD 2018
Country/Territory	France
City	Lyon
Period	24/09/18 → 27/09/18
Internet address	https://graal.ens-lyon.fr/sbac-pad/

Austrian Fields of Science 2012

102023 Supercomputing

Keywords

adaptive scheduling
concurrent workloads
runtime systems
task-based parallelism

Access to Document

10.1109/CAHPC.2018.8645869

Cite this

@inproceedings{50dbc493c3a042bbbb0ba4e0129279f4,

title = "Adaptive scheduling of collocated applications using a task-based runtime system",

abstract = "Task-based runtime systems are considered as one of the options for dealing with the challenges of upcoming parallel architectures. The greater flexibility of these runtime systems can also be used to dynamically adjust the resources allocated to the applications, adapting to the current load of the system and the progress of the applications. In our work, we have extended our implementation of the Open Community Runtime to support dynamic adjustment of execution threads. The runtimes communicate with an agent process, which collects performance data, computes thread allocation, and instructs the runtimes to make the required adjustments. We have tested our solution under different scenarios, focusing on producer-consumer applications, where the dynamic resource management was used to keep the applications in sync, improving the overall performance in some cases.",

keywords = "adaptive scheduling, concurrent workloads, runtime systems, task-based parallelism",

author = "Jiri Dokulil and Siegfried Benkner",

note = "Publisher Copyright: {\textcopyright} 2018 IEEE.; 30th International Symposium on Computer Architecture and High Performance Computing, SBAC-PAD 2018 ; Conference date: 24-09-2018 Through 27-09-2018",

year = "2018",

month = jul,

day = "2",

doi = "10.1109/CAHPC.2018.8645869",

language = "English",

pages = "41--48",

booktitle = "2018 30th International Symposium on Computer Architecture and High Performance Computing",

publisher = "IEEE",

url = "https://graal.ens-lyon.fr/sbac-pad/",

}

Dokulil, J & Benkner, S 2018, Adaptive scheduling of collocated applications using a task-based runtime system. in 2018 30th International Symposium on Computer Architecture and High Performance Computing. IEEE, pp. 41-48, 30th International Symposium on Computer Architecture and High Performance Computing, Lyon, France, 24/09/18. https://doi.org/10.1109/CAHPC.2018.8645869

TY - GEN

T1 - Adaptive scheduling of collocated applications using a task-based runtime system

AU - Dokulil, Jiri

AU - Benkner, Siegfried

PY - 2018/7/2

Y1 - 2018/7/2

N2 - Task-based runtime systems are considered as one of the options for dealing with the challenges of upcoming parallel architectures. The greater flexibility of these runtime systems can also be used to dynamically adjust the resources allocated to the applications, adapting to the current load of the system and the progress of the applications. In our work, we have extended our implementation of the Open Community Runtime to support dynamic adjustment of execution threads. The runtimes communicate with an agent process, which collects performance data, computes thread allocation, and instructs the runtimes to make the required adjustments. We have tested our solution under different scenarios, focusing on producer-consumer applications, where the dynamic resource management was used to keep the applications in sync, improving the overall performance in some cases.

AB - Task-based runtime systems are considered as one of the options for dealing with the challenges of upcoming parallel architectures. The greater flexibility of these runtime systems can also be used to dynamically adjust the resources allocated to the applications, adapting to the current load of the system and the progress of the applications. In our work, we have extended our implementation of the Open Community Runtime to support dynamic adjustment of execution threads. The runtimes communicate with an agent process, which collects performance data, computes thread allocation, and instructs the runtimes to make the required adjustments. We have tested our solution under different scenarios, focusing on producer-consumer applications, where the dynamic resource management was used to keep the applications in sync, improving the overall performance in some cases.

KW - adaptive scheduling

KW - concurrent workloads

KW - runtime systems

KW - task-based parallelism

UR - https://www.scopus.com/pages/publications/85063127875

U2 - 10.1109/CAHPC.2018.8645869

DO - 10.1109/CAHPC.2018.8645869

M3 - Contribution to proceedings

SP - 41

EP - 48

BT - 2018 30th International Symposium on Computer Architecture and High Performance Computing

PB - IEEE

T2 - 30th International Symposium on Computer Architecture and High Performance Computing

Y2 - 24 September 2018 through 27 September 2018

ER -

Adaptive scheduling of collocated applications using a task-based runtime system

Abstract

Conference

Austrian Fields of Science 2012

Keywords

Access to Document

Other files and links

Fingerprint