- 2
Pierre Ramet
Pierre RAMET - Responsable du département SATANAS (Supports et Algorithms pour les Applications Numériques Hautes Performances)
2
Documents
Affiliations actuelles
- 259761
- 104751
- 3102
- 440172
- 1124391
Identifiants chercheurs
pierre-ramet- IdRef : 189029668
-
0000-0002-6179-9819
Présentation
Ma problématique de recherche concerne les problèmes du calcul haute performance, en particulier l'algèbre linéaire creuse. Je développe la bibliothèque PaStiX ([https://solverstack.gitlabpages.inria.fr/pastix](https://solverstack.gitlabpages.inria.fr/pastix "PaStiX")) qui implémente un solveur direct creux haute performance.
---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
My interests span high-performance computing, focusing on sparse linear algebra. I am the developer of PaStiX ([https://solverstack.gitlabpages.inria.fr/pastix](https://solverstack.gitlabpages.inria.fr/pastix "PaStiX")), a high-performance sparse direct solver.
---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Domaines de recherche
Informatique [cs]
Analyse numérique [math.NA]
Compétences
Sparse Linear Algebra
Parallel Algorithms
Scientific Computing
Publications
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- 2
A task-based sparse direct solver suited for large scale hierarchical/heterogeneous architecturesSIAM Conference on Computational Science and Engineering (SIAM CSE 2015), Mar 2015, Salt Lake City, United States
Communication dans un congrès
hal-01100979v1
|
|
On the design of parallel linear solvers for large scale problemsMini-Symposium on Recent advances in matrix computations for extreme-scale computers at ICIAM'15 conference, Aug 2015, Beijing, China
Communication dans un congrès
hal-01100987v1
|