Flexible User-Defined Domain Decomposition in Kilometer-Scale E3SM Land Model Simulation

Zhuowei Gu; Dali Wang; Dawei Gao; Yunhe Feng; Qinglei Cao

doi:10.1007/978-3-031-97626-1_12

Flexible User-Defined Domain Decomposition in Kilometer-Scale E3SM Land Model Simulation

Zhuowei Gu, Dali Wang, Dawei Gao, Yunhe Feng, Qinglei Cao

Earth Systems Modeling

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

Abstract

The Energy Exascale Earth System Model (E3SM) Land Model (ELM) has been extended to kilometer-scale (km-ELM) resolutions, enabling high-fidelity simulations of terrestrial processes at 1 km × 1 km grid spacing. In ELM, domain decomposition partitions the computational domain across processors, ensuring efficient parallel execution. Currently, round-robin decomposition is applied, providing a straightforward way to distribute computational workload. As ELM continues evolving at the kilometer-scale (km-scale), particularly with integrating lateral flow modeling, decomposition strategies must also account for the increased workload and data movement. This paper introduces a flexible user-defined domain decomposition framework, allowing users to customize domain partitioning based on application requirements. The impact of different decomposition strategies is evaluated across various applications concerning computation, communication, and I/O. Results demonstrate that while 1D partitioning yields superior I/O performance, k-nearest neighbors (KNN) clustering effectively reduces inter-process communication overhead. This study lays the groundwork for scalable partitioning in large-scale land surface simulations, enhancing next-generation Earth system modeling.

Original language	English
Title of host publication	Computational Science – ICCS 2025 - 25th International Conference, Proceedings
Editors	Michael H. Lees, Wentong Cai, Siew Ann Cheong, Yi Su, David Abramson, Jack J. Dongarra, Peter M. A. Sloot
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	164-178
Number of pages	15
ISBN (Print)	9783031976254
DOIs	https://doi.org/10.1007/978-3-031-97626-1_12
State	Published - 2025
Event	25th International Conference on Computational Science, ICCS 2025 - Singapore, Singapore Duration: Jul 7 2025 → Jul 9 2025

Publication series

Name	Lecture Notes in Computer Science
Volume	15903 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	25th International Conference on Computational Science, ICCS 2025
Country/Territory	Singapore
City	Singapore
Period	07/7/25 → 07/9/25

Funding

This research is supported as part of the Energy Exascale Earth System Model (E3SM) project funded by DOE’s Office of Science, Office of Biological and Environmental Research. This project leverages the Daymet data product funded by NASA and uses resources of the Oak Ridge Leadership Computing Facility, supported by DOE’s Office of Science under Contract No. DE-AC05-00OR22725.

Keywords

Domain Decomposition
E3SM Land Model (ELM)
Earth System Modeling
I/O
Kilometer-Scale ELM
Load Balancing

Access to Document

10.1007/978-3-031-97626-1_12

Cite this

Gu, Z., Wang, D., Gao, D., Feng, Y., & Cao, Q. (2025). Flexible User-Defined Domain Decomposition in Kilometer-Scale E3SM Land Model Simulation. In M. H. Lees, W. Cai, S. A. Cheong, Y. Su, D. Abramson, J. J. Dongarra, & P. M. A. Sloot (Eds.), Computational Science – ICCS 2025 - 25th International Conference, Proceedings (pp. 164-178). (Lecture Notes in Computer Science; Vol. 15903 LNCS). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-031-97626-1_12

Gu, Zhuowei ; Wang, Dali ; Gao, Dawei et al. / Flexible User-Defined Domain Decomposition in Kilometer-Scale E3SM Land Model Simulation. Computational Science – ICCS 2025 - 25th International Conference, Proceedings. editor / Michael H. Lees ; Wentong Cai ; Siew Ann Cheong ; Yi Su ; David Abramson ; Jack J. Dongarra ; Peter M. A. Sloot. Springer Science and Business Media Deutschland GmbH, 2025. pp. 164-178 (Lecture Notes in Computer Science).

@inproceedings{d171d754982a44aa8faf5a171df67f60,

title = "Flexible User-Defined Domain Decomposition in Kilometer-Scale E3SM Land Model Simulation",

abstract = "The Energy Exascale Earth System Model (E3SM) Land Model (ELM) has been extended to kilometer-scale (km-ELM) resolutions, enabling high-fidelity simulations of terrestrial processes at 1 km × 1 km grid spacing. In ELM, domain decomposition partitions the computational domain across processors, ensuring efficient parallel execution. Currently, round-robin decomposition is applied, providing a straightforward way to distribute computational workload. As ELM continues evolving at the kilometer-scale (km-scale), particularly with integrating lateral flow modeling, decomposition strategies must also account for the increased workload and data movement. This paper introduces a flexible user-defined domain decomposition framework, allowing users to customize domain partitioning based on application requirements. The impact of different decomposition strategies is evaluated across various applications concerning computation, communication, and I/O. Results demonstrate that while 1D partitioning yields superior I/O performance, k-nearest neighbors (KNN) clustering effectively reduces inter-process communication overhead. This study lays the groundwork for scalable partitioning in large-scale land surface simulations, enhancing next-generation Earth system modeling.",

keywords = "Domain Decomposition, E3SM Land Model (ELM), Earth System Modeling, I/O, Kilometer-Scale ELM, Load Balancing",

author = "Zhuowei Gu and Dali Wang and Dawei Gao and Yunhe Feng and Qinglei Cao",

note = "Publisher Copyright: {\textcopyright} The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.; 25th International Conference on Computational Science, ICCS 2025 ; Conference date: 07-07-2025 Through 09-07-2025",

year = "2025",

doi = "10.1007/978-3-031-97626-1\_12",

language = "English",

isbn = "9783031976254",

series = "Lecture Notes in Computer Science",

publisher = "Springer Science and Business Media Deutschland GmbH",

pages = "164--178",

editor = "Lees, \{Michael H.\} and Wentong Cai and Cheong, \{Siew Ann\} and Yi Su and David Abramson and Dongarra, \{Jack J.\} and Sloot, \{Peter M. A.\}",

booktitle = "Computational Science – ICCS 2025 - 25th International Conference, Proceedings",

}

Gu, Z, Wang, D, Gao, D, Feng, Y & Cao, Q 2025, Flexible User-Defined Domain Decomposition in Kilometer-Scale E3SM Land Model Simulation. in MH Lees, W Cai, SA Cheong, Y Su, D Abramson, JJ Dongarra & PMA Sloot (eds), Computational Science – ICCS 2025 - 25th International Conference, Proceedings. Lecture Notes in Computer Science, vol. 15903 LNCS, Springer Science and Business Media Deutschland GmbH, pp. 164-178, 25th International Conference on Computational Science, ICCS 2025, Singapore, Singapore, 07/7/25. https://doi.org/10.1007/978-3-031-97626-1_12

Flexible User-Defined Domain Decomposition in Kilometer-Scale E3SM Land Model Simulation. / Gu, Zhuowei; Wang, Dali; Gao, Dawei et al.
Computational Science – ICCS 2025 - 25th International Conference, Proceedings. ed. / Michael H. Lees; Wentong Cai; Siew Ann Cheong; Yi Su; David Abramson; Jack J. Dongarra; Peter M. A. Sloot. Springer Science and Business Media Deutschland GmbH, 2025. p. 164-178 (Lecture Notes in Computer Science; Vol. 15903 LNCS).

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

TY - GEN

T1 - Flexible User-Defined Domain Decomposition in Kilometer-Scale E3SM Land Model Simulation

AU - Gu, Zhuowei

AU - Wang, Dali

AU - Gao, Dawei

AU - Feng, Yunhe

AU - Cao, Qinglei

N1 - Publisher Copyright: © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

PY - 2025

Y1 - 2025

N2 - The Energy Exascale Earth System Model (E3SM) Land Model (ELM) has been extended to kilometer-scale (km-ELM) resolutions, enabling high-fidelity simulations of terrestrial processes at 1 km × 1 km grid spacing. In ELM, domain decomposition partitions the computational domain across processors, ensuring efficient parallel execution. Currently, round-robin decomposition is applied, providing a straightforward way to distribute computational workload. As ELM continues evolving at the kilometer-scale (km-scale), particularly with integrating lateral flow modeling, decomposition strategies must also account for the increased workload and data movement. This paper introduces a flexible user-defined domain decomposition framework, allowing users to customize domain partitioning based on application requirements. The impact of different decomposition strategies is evaluated across various applications concerning computation, communication, and I/O. Results demonstrate that while 1D partitioning yields superior I/O performance, k-nearest neighbors (KNN) clustering effectively reduces inter-process communication overhead. This study lays the groundwork for scalable partitioning in large-scale land surface simulations, enhancing next-generation Earth system modeling.

AB - The Energy Exascale Earth System Model (E3SM) Land Model (ELM) has been extended to kilometer-scale (km-ELM) resolutions, enabling high-fidelity simulations of terrestrial processes at 1 km × 1 km grid spacing. In ELM, domain decomposition partitions the computational domain across processors, ensuring efficient parallel execution. Currently, round-robin decomposition is applied, providing a straightforward way to distribute computational workload. As ELM continues evolving at the kilometer-scale (km-scale), particularly with integrating lateral flow modeling, decomposition strategies must also account for the increased workload and data movement. This paper introduces a flexible user-defined domain decomposition framework, allowing users to customize domain partitioning based on application requirements. The impact of different decomposition strategies is evaluated across various applications concerning computation, communication, and I/O. Results demonstrate that while 1D partitioning yields superior I/O performance, k-nearest neighbors (KNN) clustering effectively reduces inter-process communication overhead. This study lays the groundwork for scalable partitioning in large-scale land surface simulations, enhancing next-generation Earth system modeling.

KW - Domain Decomposition

KW - E3SM Land Model (ELM)

KW - Earth System Modeling

KW - I/O

KW - Kilometer-Scale ELM

KW - Load Balancing

UR - https://www.scopus.com/pages/publications/105010835272

U2 - 10.1007/978-3-031-97626-1_12

DO - 10.1007/978-3-031-97626-1_12

M3 - Conference contribution

AN - SCOPUS:105010835272

SN - 9783031976254

T3 - Lecture Notes in Computer Science

SP - 164

EP - 178

BT - Computational Science – ICCS 2025 - 25th International Conference, Proceedings

A2 - Lees, Michael H.

A2 - Cai, Wentong

A2 - Cheong, Siew Ann

A2 - Su, Yi

A2 - Abramson, David

A2 - Dongarra, Jack J.

A2 - Sloot, Peter M. A.

PB - Springer Science and Business Media Deutschland GmbH

T2 - 25th International Conference on Computational Science, ICCS 2025

Y2 - 7 July 2025 through 9 July 2025

ER -

Gu Z, Wang D, Gao D, Feng Y, Cao Q. Flexible User-Defined Domain Decomposition in Kilometer-Scale E3SM Land Model Simulation. In Lees MH, Cai W, Cheong SA, Su Y, Abramson D, Dongarra JJ, Sloot PMA, editors, Computational Science – ICCS 2025 - 25th International Conference, Proceedings. Springer Science and Business Media Deutschland GmbH. 2025. p. 164-178. (Lecture Notes in Computer Science). doi: 10.1007/978-3-031-97626-1_12