Multilevel communication optimal LU and QR factorizations for hierarchical platforms

Laura Grigori 1 Mathias Jacquelin 1 Amal Khabou 1
1 ALPINES - Algorithms and parallel tools for integrated numerical simulations
LJLL - Laboratoire Jacques-Louis Lions, Inria Paris-Rocquencourt, Institut National des Sciences Mathématiques et de leurs Interactions
Abstract : This study focuses on the performance of two classical dense linear algebra algorithms, the LU and the QR factorizations, on multilevel hierarchical platforms. We first introduce a new model called Hierarchical Cluster Platform (HCP), encapsulating the characteristics of such platforms. The focus is set on reducing the communication requirements of studied algorithms at each level of the hierarchy. Lower bounds on communications are therefore extended with respect to the HCP model. We then introduce multilevel LU and QR algorithms tailored for those platforms, and provide a detailed performance analysis. We also provide a set of numerical experiments and performance predictions demonstrating the need for such algorithms on large platforms.
Document type :
Preprints, Working Papers, ...
Complete list of metadatas

Cited literature [15 references]  Display  Hide  Download

https://hal.inria.fr/hal-00803718
Contributor : Mathias Jacquelin <>
Submitted on : Friday, March 22, 2013 - 4:08:01 PM
Last modification on : Wednesday, May 15, 2019 - 3:40:45 AM
Long-term archiving on : Sunday, April 2, 2017 - 6:24:17 PM

Files

RR-8270.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-00803718, version 1
  • ARXIV : 1303.5837

Citation

Laura Grigori, Mathias Jacquelin, Amal Khabou. Multilevel communication optimal LU and QR factorizations for hierarchical platforms. 2013. ⟨hal-00803718⟩

Share

Metrics

Record views

300

Files downloads

811