Toward a Multi-Target Linear Algebra Library for GPUs and Multicore CPUs

Laurent Plagne 1 Wilfried Kirschenmann 1, 2 Stéphane Vialle 2, 3
2 ALGORILLE - Algorithms for the Grid
INRIA Lorraine, LORIA - Laboratoire Lorrain de Recherche en Informatique et ses Applications
Abstract : Highly structured sparse matrices arise frequently from numerical discretizations of partial differential equations. Legolas++ is a C++ generic library designed for describing and manipulating such multi-level blocked matrices with the corresponding blocked vectors and algorithms. Legolas++ allows a very detailed description of the linear systems to be solved that can be used to generate efficient parallel implementations. We are working on enlarging the scope of Legolas++ from multi- core target to GPUs and cluster of GPUs.
Complete list of metadatas

https://hal-supelec.archives-ouvertes.fr/hal-00525899
Contributor : Sébastien van Luchene <>
Submitted on : Wednesday, October 13, 2010 - 10:53:23 AM
Last modification on : Thursday, March 29, 2018 - 11:06:04 AM

Identifiers

  • HAL Id : hal-00525899, version 1

Collections

Citation

Laurent Plagne, Wilfried Kirschenmann, Stéphane Vialle. Toward a Multi-Target Linear Algebra Library for GPUs and Multicore CPUs. SIAM Conference on Parallel Processing for Scientific Computing (PP10), Feb 2010, Seattle, Washington, United States. pp.MS55. ⟨hal-00525899⟩

Share

Metrics

Record views

451