The Unassembled Hyper Matrix (UHM) solver is developed for the efficient solution of large sparse linear systems of equations arising from Finite Element Methods (FEM). The efficiency of UHM solver comes from the direct use of information from an application instead of analyzing a given sparse matrix only. While most sparse direct solvers adopt a black box approach: a sparse linear system of equations is given as input and the solution is returned as output, UHM solver focuses on the problem domain information and the interfaces them. Especially UHM solver outperforms the other sparse direct solvers for the problems that use high order FE approximation. This solver is built on the FLAME library and follows its methodologies.
Keywords for this software
References in zbMATH (referenced in 7 articles , 1 standard article )
Showing results 1 to 7 of 7.
- Fialko, Sergiy: Parallel finite element solver for multi-core computers with shared memory (2021)
- Davis, Timothy A.; Duff, Iain S.; Nakov, Stojce: Design and implementation of a parallel Markowitz threshold algorithm (2020)
- Agullo, Emmanuel; Buttari, Alfredo; Guermouche, Abdou; Lopez, Florent: Implementing multifrontal sparse solvers for multicore architectures with sequential task flow runtime systems (2016)
- Ghysels, Pieter; Li, Xiaoye S.; Rouet, François-Henry; Williams, Samuel; Napov, Artem: An efficient multicore implementation of a novel HSS-structured multifrontal solver using randomized sampling (2016)
- Hogg, Jonathan D.; Ovtchinnikov, Evgueni; Scott, Jennifer A.: A sparse symmetric indefinite direct solver for GPU architectures (2016)
- Fialko, Sergiy: Parallel direct solver for solving systems of linear equations resulting from finite element method on multi-core desktops and workstations (2015)
- Kim, Kyungjoo; Eijkhout, Victor: A parallel sparse direct solver via hierarchical DAG scheduling (2014)