Enhancing Sparse Direct Solver Scalability Through Runtime System Automatic Data Partition
Uloženo v:
| Název: | Enhancing Sparse Direct Solver Scalability Through Runtime System Automatic Data Partition |
|---|---|
| Autoři: | Lisito, Alycia, Faverge, Mathieu, Pichon, Grégoire, Ramet, Pierre |
| Přispěvatelé: | Lisito, Alycia |
| Zdroj: | Lecture Notes in Computer Science ISBN: 9783031617621 |
| Informace o vydavateli: | Springer Nature Switzerland, 2024. |
| Rok vydání: | 2024 |
| Témata: | Task scheduling, [INFO.INFO-DC] Computer Science [cs]/Distributed, Parallel, and Cluster Computing [cs.DC], 1. No poverty, Runtime systems, Sparse Direct Solver, Data partitioning |
| Popis: | With the ever-growing number of cores per node, it is critical for runtime systems and applications to adapt the task granularity to scale on recent architectures. Among applications, sparse direct solvers are a time-consuming step and the task granularity is rarely adapted to large many-core systems. In this paper, we investigate the use of runtimesystems to automatically partition tasks in order to achieve more parallelism and refine the task granularity. Experiments are conducted on the new version of the PaStiX solver, which has been completely rewritten to better integrate modern task-based runtime systems. The results demonstrate the increase in scalability achieved by the solver thanks to the adaptive task granularity provided by the StarPU runtime system. |
| Druh dokumentu: | Part of book or chapter of book Conference object |
| Popis souboru: | application/pdf |
| Jazyk: | English |
| DOI: | 10.1007/978-3-031-61763-8_10 |
| Přístupová URL adresa: | https://inria.hal.science/hal-04527103v1 https://inria.hal.science/hal-04527103v1/document https://doi.org/10.1007/978-3-031-61763-8_10 |
| Rights: | Springer Nature TDM CC BY |
| Přístupové číslo: | edsair.doi.dedup.....d9737391054713158edb8e446c43dfa1 |
| Databáze: | OpenAIRE |
Nájsť tento článok vo Web of Science | Abstrakt: | With the ever-growing number of cores per node, it is critical for runtime systems and applications to adapt the task granularity to scale on recent architectures. Among applications, sparse direct solvers are a time-consuming step and the task granularity is rarely adapted to large many-core systems. In this paper, we investigate the use of runtimesystems to automatically partition tasks in order to achieve more parallelism and refine the task granularity. Experiments are conducted on the new version of the PaStiX solver, which has been completely rewritten to better integrate modern task-based runtime systems. The results demonstrate the increase in scalability achieved by the solver thanks to the adaptive task granularity provided by the StarPU runtime system. |
|---|---|
| DOI: | 10.1007/978-3-031-61763-8_10 |