Static LU Decomposition on Heterogeneous Platforms

Olivier Beaumont 1 Arnaud Legrand 1 Fabrice Rastello 1 Yves Robert 1
1 REMAP - Regularity and massive parallel computing
Inria Grenoble - Rhône-Alpes, LIP - Laboratoire de l'Informatique du Parallélisme
Abstract : In this paper, we deal with algorithmic issues on heterogeneous platforms. We concentrate on dense linear algebra kernels, such as matrix multiplication or LU decomposition. Block cyclic distribution techniques used in ScaLAPACK are no longer sufficient to balance the load among processors running at different speeds. The main result of this paper is to provide a static data distribution scheme that leads to an asymptotically perfect load balancing for LU decomposition, thereby providing solid foundations toward the design of a cluster-oriented version of ScaLAPACK.
Complete list of metadatas

https://hal.inria.fr/hal-00856641
Contributor : Equipe Roma <>
Submitted on : Monday, September 2, 2013 - 10:21:56 AM
Last modification on : Friday, April 20, 2018 - 3:44:24 PM

Identifiers

  • HAL Id : hal-00856641, version 1

Collections

Citation

Olivier Beaumont, Arnaud Legrand, Fabrice Rastello, Yves Robert. Static LU Decomposition on Heterogeneous Platforms. [Research Report] 2000-44, 2000. ⟨hal-00856641⟩

Share

Metrics

Record views

276