Detail publikace

Cache Efficient Implementation for Block Matrix Operations

POLOK, L. ILA, V. SMRŽ, P.

Originální název

Cache Efficient Implementation for Block Matrix Operations

Anglický název

Cache Efficient Implementation for Block Matrix Operations

Jazyk

en

Originální abstrakt

Efficiently manipulating and operating on block matrices can be beneficial in many applications, among others those involving iteratively solving nonlinear systems. These types of problems consist of repeatedly assembling and solving sparse linear systems. In the case of very large systems, without a careful manipulation of the corresponding matrices, solving can become very time consuming. This paper proposes a memory storage scheme convenient for both, numeric and structural matrix modification and, at the same time, allowing efficient arithmetic operation. This scheme was used in the implementation of a simple BLAS-like library. The advantage of the new scheme is demonstrated through exhaustive tests on the popular University of Florida Sparse Matrix Collection. Furthermore, this library was used in solving several nonlinear graph optimization problems.

Anglický abstrakt

Efficiently manipulating and operating on block matrices can be beneficial in many applications, among others those involving iteratively solving nonlinear systems. These types of problems consist of repeatedly assembling and solving sparse linear systems. In the case of very large systems, without a careful manipulation of the corresponding matrices, solving can become very time consuming. This paper proposes a memory storage scheme convenient for both, numeric and structural matrix modification and, at the same time, allowing efficient arithmetic operation. This scheme was used in the implementation of a simple BLAS-like library. The advantage of the new scheme is demonstrated through exhaustive tests on the popular University of Florida Sparse Matrix Collection. Furthermore, this library was used in solving several nonlinear graph optimization problems.

Dokumenty

BibTex


@inproceedings{BUT103462,
  author="Lukáš {Polok} and Viorela Simona {Ila} and Pavel {Smrž}",
  title="Cache Efficient Implementation for Block Matrix Operations",
  annote="Efficiently manipulating and operating on block matrices can be beneficial in
many applications, among others those involving iteratively solving nonlinear
systems. These types of problems consist of repeatedly assembling and solving
sparse linear systems. In the case of very large systems, without a careful
manipulation of the corresponding matrices, solving can become very time
consuming.

This paper proposes a memory storage scheme convenient for both, numeric and
structural matrix modification and, at the same time, allowing efficient
arithmetic operation. This scheme was used in the implementation of a simple
BLAS-like library. The advantage of the new scheme is demonstrated through
exhaustive tests on the popular University of Florida Sparse Matrix Collection.
Furthermore, this library was used in solving several nonlinear graph
optimization problems.",
  address="Association for Computing Machinery",
  booktitle="Proceedings of the 21st High Performance Computing Symposium (HPC'13)",
  chapter="103462",
  edition="NEUVEDEN",
  howpublished="online",
  institution="Association for Computing Machinery",
  year="2013",
  month="april",
  pages="698--706",
  publisher="Association for Computing Machinery",
  type="conference paper"
}