Towards efficient high-resolution simulations of dust storm formation over large areas in North Africa using HPC

This study presents the results of scalability tests of the WRF-Chem model on the MareNostrum 5 GPP supercomputing system, aimed at enabling high-resolution simulations of dust storm formation across large regions of North Africa. Simulations were conducted with three nested domains, reaching a 2 km horizontal resolution in the innermost domain. Two large areas of different size were evaluated. For the smaller domain, the optimal wall-clock time was achieved using approximately 2,000 cores, resulting in a simulation-to-wall-clock time ratio of 2:1. However, parallel efficiency declined significantly beyond ~1,200 cores, indicating limited scalability at higher core counts. In contrast, simulations over the larger domain maintained high efficiency between 1,680 and 2,016 cores, even outperforming the 2,500-core configuration in normalized cost. These findings highlight the importance of balancing runtime and resource efficiency. Future work should investigate which components of WRF-Chem dominate computational cost to better exploit HPC resources and guide model optimization

Authors

Institutions

Publication Details

Journal
Procedia Computer Science
Published
2025-01-01
DOI
https://doi.org/10.1016/j.procs.2025.08.250
Primary Topic
Atmospheric aerosols and clouds
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

Towards efficient high-resolution simulations of dust storm formation over large areas in North Africa using HPC

J.A.G. Orza, Samira Karbasi
Procedia Computer Science
Atmospheric aerosols and clouds
article

Towards efficient high-resolution simulations of dust storm formation over large areas in North Africa using HPC

J.A.G. Orza, Samira Karbasi
article en

Abstract

This study presents the results of scalability tests of the WRF-Chem model on the MareNostrum 5 GPP supercomputing system, aimed at enabling high-resolution simulations of dust storm formation across large regions of North Africa. Simulations were conducted with three nested domains, reaching a 2 km horizontal resolution in the innermost domain. Two large areas of different size were evaluated. For the smaller domain, the optimal wall-clock time was achieved using approximately 2,000 cores, resulting in a simulation-to-wall-clock time ratio of 2:1. However, parallel efficiency declined significantly beyond ~1,200 cores, indicating limited scalability at higher core counts. In contrast, simulations over the larger domain maintained high efficiency between 1,680 and 2,016 cores, even outperforming the 2,500-core configuration in normalized cost. These findings highlight the importance of balancing runtime and resource efficiency. Future work should investigate which components of WRF-Chem dominate computational cost to better exploit HPC resources and guide model optimization

Procedia Computer ScienceVol. 267
Universitat de Miguel Hernández d'Elx (ES)
Openalex Percentile: Top 75%
Atmospheric aerosols and clouds
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.