On the use of low rank approximations for sparse direct solvers - Inria - Institut national de recherche en sciences et technologies du numérique Access content directly
Conference Papers Year : 2016

On the use of low rank approximations for sparse direct solvers

Abstract

In this talk, we describe a preliminary fast direct solver using HODLR library to compress large blocks appearing in the symbolic structure of the PaStiX sparse direct solver. We will present our general strategy before analyzing the practical gains in terms of memory and floating point operations with respect to a theoretical study of the problem. Finally, we will discuss the impact of reordering technic to enhance the low-rank compression.
No file

Dates and versions

hal-01421376 , version 1 (22-12-2016)

Identifiers

  • HAL Id : hal-01421376 , version 1

Cite

Grégoire Pichon, Eric Darve, Mathieu Faverge, Pierre Ramet, Jean Roman. On the use of low rank approximations for sparse direct solvers. SIAM Annual Meeting (AN'16), Jul 2016, Boston, United States. ⟨hal-01421376⟩
242 View
0 Download

Share

Gmail Facebook X LinkedIn More