Using MPIs Non-Blocking Allreduce for Health Checks in Dynamic Simulations

Gericke, Jana and Klimach, Harald Günther and Ebrahimi Pour, Neda and Roller, Sabine (2023) Using MPIs Non-Blocking Allreduce for Health Checks in Dynamic Simulations. In: Parallel and Distributed Computing, Applications and Technologies, PDCAT 2023, 1112. The 24th International Conference on Parallel and Distributed Computing, Applications and Technologies (PDCAT 2023), 2023-08-16 - 2023-08-18, Jeju, Korea. doi: 10.1007/978-981-99-8211-0_3. ISBN 978-981998210-3. ISSN 1876-1100.

This is the latest version of this item.

PDF - Only accessible within DLR
1MB

Abstract

Large-scale simulations often require frequent checks on global conditions that are not directly needed for the computation itself. While such health checks are not an integral part of the numerical algorithm, they serve an important role in controlling and coordinating the simulation. In distributed parallel computations their required communication may negatively impact the actual parallel computation and result in unnecessary synchronization points. By using a nonblocking reduction these synchronization requirements can be loosened and the impact on the actual computation minimized. Further, it enables us to shift the communication into the background and progress it during the MPI calls that are done during the computation anyway. This approach delays the decision to be made in response to those health checks. But as it is not vital for the correct computation itself such a delay is usually tolerable and could offer a more robust scaling to large process counts.

Item URL in elib:

https://elib.dlr.de/198232/

Document Type:

Conference or Workshop Item (Speech)

Title:

Using MPIs Non-Blocking Allreduce for Health Checks in Dynamic Simulations

Authors:

Authors	Institution or Email of Authors	Author's ORCID iD	ORCID Put Code
Gericke, Jana	UNSPECIFIED	https://orcid.org/0000-0003-0322-2197	149817790
Klimach, Harald Günther	UNSPECIFIED	https://orcid.org/0000-0002-6054-5681	149817791
Ebrahimi Pour, Neda	UNSPECIFIED	https://orcid.org/0000-0002-8167-7456	UNSPECIFIED
Roller, Sabine	UNSPECIFIED	https://orcid.org/0000-0003-1483-8456	149817792

Date:

17 August 2023

Journal or Publication Title:

Parallel and Distributed Computing, Applications and Technologies, PDCAT 2023

Refereed publication:

Yes

Open Access:

Gold Open Access:

In SCOPUS:

Yes

In ISI Web of Science:

Volume:

1112

DOI:

10.1007/978-981-99-8211-0_3

Series Name:

Lecture Notes in Electrical Engineering

ISSN:

1876-1100

ISBN:

978-981998210-3

Status:

Published

Keywords:

MPI, Non-Blocking Collectives, Dynamic Simulations

Event Title:

The 24th International Conference on Parallel and Distributed Computing, Applications and Technologies (PDCAT 2023)

Event Location:

Jeju, Korea

Event Type:

international Conference

Event Start Date:

16 August 2023

Event End Date:

18 August 2023

Organizer:

BIC, PDCAT & KIPS CSWRG

HGF - Research field:

Aeronautics, Space and Transport

HGF - Program:

Aeronautics

HGF - Program Themes:

other

DLR - Research area:

Aeronautics

DLR - Program:

L - no assignment

DLR - Research theme (Project):

L - no assignment

Location:

Dresden

Institutes and Institutions:

Institute of Software Methods for Product Virtualization

Deposited By:

Gericke, Jana

Deposited On:

02 Nov 2023 15:29

Last Modified:

24 Apr 2024 20:58

Available Versions of this Item

Using MPIs Non-Blocking Allreduce for Health Checks in Dynamic Simulations. (deposited 02 Nov 2023 15:29) [Currently Displayed]

Repository Staff Only: item control page