Exploiting nested task-parallelism in the $\mathcal{H}-LU$ factorization
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2019
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866913509843927040 |
|---|---|
| author | Carratalá-Sáez, Rocío Christophersen, Sven Aliaga, José I. Beltran, Vicenç Börm, Steffen Quintana-Ortí, Enrique S. |
| author_facet | Carratalá-Sáez, Rocío Christophersen, Sven Aliaga, José I. Beltran, Vicenç Börm, Steffen Quintana-Ortí, Enrique S. |
| contents | We address the parallelization of the LU factorization of hierarchical matrices ($\mathcal{H}$-matrices) arising from boundary element methods. Our approach exploits task-parallelism via the OmpSs programming model and runtime, which discovers the data-flow parallelism intrinsic to the operation at execution time, via the analysis of data dependencies based on the memory addresses of the tasks' operands. This is especially challenging for $\mathcal{H}$-matrices, as the structures containing the data vary in dimension during the execution. We tackle this issue by decoupling the data structure from that used to detect dependencies. Furthermore, we leverage the support for weak operands and early release of dependencies, recently introduced in OmpSs-2, to accelerate the execution of parallel codes with nested task-parallelism and fine-grain tasks. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_1906_00874 |
| institution | arXiv |
| publishDate | 2019 |
| record_format | arxiv |
| spellingShingle | Exploiting nested task-parallelism in the $\mathcal{H}-LU$ factorization Carratalá-Sáez, Rocío Christophersen, Sven Aliaga, José I. Beltran, Vicenç Börm, Steffen Quintana-Ortí, Enrique S. Mathematical Software Distributed, Parallel, and Cluster Computing Numerical Analysis 68W10, 65N38, 65F05 We address the parallelization of the LU factorization of hierarchical matrices ($\mathcal{H}$-matrices) arising from boundary element methods. Our approach exploits task-parallelism via the OmpSs programming model and runtime, which discovers the data-flow parallelism intrinsic to the operation at execution time, via the analysis of data dependencies based on the memory addresses of the tasks' operands. This is especially challenging for $\mathcal{H}$-matrices, as the structures containing the data vary in dimension during the execution. We tackle this issue by decoupling the data structure from that used to detect dependencies. Furthermore, we leverage the support for weak operands and early release of dependencies, recently introduced in OmpSs-2, to accelerate the execution of parallel codes with nested task-parallelism and fine-grain tasks. |
| title | Exploiting nested task-parallelism in the $\mathcal{H}-LU$ factorization |
| topic | Mathematical Software Distributed, Parallel, and Cluster Computing Numerical Analysis 68W10, 65N38, 65F05 |
| url | https://arxiv.org/abs/1906.00874 |