Exploiting Parameterized Task-graph in Sparse Direct Solvers

Abstract : Task-based programming models have been widely studied in the context of dense linear algebra, but remains less studied for the more complex sparse solvers. In this talk, we will present the use of two different programming models: Sequential Task Flow from StarPU, and Parameterized Task Graph from PaRSEC to parallelize the factorization step of the PaStiX sparse direct solver. We will present how those programming models have been used to integrate more complex and finer parallelism to take into account new architectures with many computational units. Efficiency of such solutions on homogeneous and heterogeneous architectures with a spectrum of matrices from different applications will be shown. We will also present how such solutions enable, without extra cost to the programmer, better performance on irregular computations such as in the block low-rank implementation of the solver.
Complete list of metadatas

https://hal.inria.fr/hal-01956963
Contributor : Mathieu Faverge <>
Submitted on : Sunday, December 16, 2018 - 11:27:41 PM
Last modification on : Friday, October 25, 2019 - 3:08:02 PM

Identifiers

  • HAL Id : hal-01956963, version 1

Citation

Mathieu Faverge, Grégoire Pichon, Pierre Ramet. Exploiting Parameterized Task-graph in Sparse Direct Solvers. SIAM Conference on Computational Science and Engineering (CSE19), Feb 2019, Spokane, United States. ⟨hal-01956963⟩

Share

Metrics

Record views

59