Investigating Matrix Repartitioning to Address the Over- and Undersubscription Challenge for a GPU-based CFD Solver

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Olenik, Gregor, Koch, Marcel, Anzt, Hartwig
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866909833331998720
author Olenik, Gregor
Koch, Marcel
Anzt, Hartwig
author_facet Olenik, Gregor
Koch, Marcel
Anzt, Hartwig
contents Modern high-performance computing (HPC) increasingly relies on GPUs, but integrating GPU acceleration into complex scientific frameworks like OpenFOAM remains a challenge. Existing approaches either fully refactor the codebase or use plugin-based GPU solvers, each facing trade-offs between performance and development effort. In this work, we address the limitations of plugin-based GPU acceleration in OpenFOAM by proposing a repartitioning strategy that better balances CPU matrix assembly and GPU-based linear solves. We present a detailed computational model, describe a novel matrix repartitioning and update procedure, and evaluate its performance on large-scale CFD simulations. Our results show that the proposed method significantly mitigates oversubscription issues, improving solver performance and resource utilization in heterogeneous CPU-GPU environments.
format Preprint
id arxiv_https___arxiv_org_abs_2510_08536
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Investigating Matrix Repartitioning to Address the Over- and Undersubscription Challenge for a GPU-based CFD Solver
Olenik, Gregor
Koch, Marcel
Anzt, Hartwig
Distributed, Parallel, and Cluster Computing
Software Engineering
Modern high-performance computing (HPC) increasingly relies on GPUs, but integrating GPU acceleration into complex scientific frameworks like OpenFOAM remains a challenge. Existing approaches either fully refactor the codebase or use plugin-based GPU solvers, each facing trade-offs between performance and development effort. In this work, we address the limitations of plugin-based GPU acceleration in OpenFOAM by proposing a repartitioning strategy that better balances CPU matrix assembly and GPU-based linear solves. We present a detailed computational model, describe a novel matrix repartitioning and update procedure, and evaluate its performance on large-scale CFD simulations. Our results show that the proposed method significantly mitigates oversubscription issues, improving solver performance and resource utilization in heterogeneous CPU-GPU environments.
title Investigating Matrix Repartitioning to Address the Over- and Undersubscription Challenge for a GPU-based CFD Solver
topic Distributed, Parallel, and Cluster Computing
Software Engineering
url https://arxiv.org/abs/2510.08536