Show simple item record

dc.contributor.authorBernabeu, S.
dc.contributor.authorPuzyrev, Volodymyr
dc.contributor.authorHanzich, M.
dc.contributor.authorFernandez, S.
dc.date.accessioned2017-01-30T12:08:07Z
dc.date.available2017-01-30T12:08:07Z
dc.date.created2016-10-10T19:30:20Z
dc.date.issued2015
dc.identifier.citationBernabeu, S. and Puzyrev, V. and Hanzich, M. and Fernandez, S. 2015. Efficient sparse matrix-vector multiplication for geophysical electromagnetic codes on Xeon Phi coprocessors, in Proceedings of the Second EAGE Workshop on High Performance Computing for Upstream, Sep 13-16 2015, pp. 61-65. Dubai, UAE: European Association of Geoscientists & Engineers (EAGE).
dc.identifier.urihttp://hdl.handle.net/20.500.11937/18477
dc.identifier.doi10.3997/2214-4609.201414033
dc.description.abstract

Sparse matrix-vector multiplication (spMV) is a fundamental building block of iterative solvers in many scientific applications. spMV is known to perform poorly in modern processors due to excessive pressure over the memory system, overhead of irregular memory accesses and load imbalance due to non-uniform matrix structures. Achieving higher performance requires taking advantage of the features of the matrix and choosing the right sparse storage format to better exploit the target architecture. In this paper we describe an efficient spMV for geophysical electromagnetic simulations on Intel Xeon Phi coprocessors. The unique features of the matrix resulting from electromagnetic problems make it hard to handle with classical sparse storage formats. We propose a matrix decomposition and a tuned storage format that obtains a 4.13x performance improvement over the optimized CSR spMV kernel on Xeon Phi coprocessors.

dc.titleEfficient sparse matrix-vector multiplication for geophysical electromagnetic codes on Xeon Phi coprocessors
dc.typeConference Paper
dcterms.source.startPage61
dcterms.source.endPage65
dcterms.source.title2nd EAGE Workshop on High Performance Computing for Upstream
dcterms.source.series2nd EAGE Workshop on High Performance Computing for Upstream
dcterms.source.isbn9781510814165
curtin.departmentDepartment of Applied Geology
curtin.accessStatusFulltext not available


Files in this item

FilesSizeFormatView

There are no files associated with this item.

This item appears in the following Collection(s)

Show simple item record