User Tools

Site Tools


dbcsr

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revision Previous revision
Next revision
Previous revision
dbcsr [2018/01/18 22:31]
oschuett
dbcsr [2018/07/17 09:48] (current)
alazzaro
Line 2: Line 2:
  
 libDBCSR (or DBCSR for short) is a sparse matrix library designed to efficiently perform sparse matrix matrix multiplication,​ among other operations. It is MPI and OpenMP parallel, and can exploit accelerators. It is developed as part of [[ http://​www.cp2k.org/​ | CP2K ]], where it provides core functionality for [[ doi>​10.1021/​ct200897x | linear scaling electronic structure theory ]]. A general overview of the library [[ doi>​10.1016/​j.parco.2014.03.012 | has been published ]]. A discussion of recent developments,​ in particular GPU work, [[ doi>​10.1002/​9781118670712.ch8 | has appeared as a chapter]] in '​Electronic Structure Calculations on Graphics Processing Units',​ John Wiley and Sons, ISBN 9781118661789,​ and is available as a {{ :​gpu_book_chapter_submitted.pdf |preprint }}. The use of one-sided MPI and a 2.5D algorithm to reduce communication is shown to be effective for sparse matrix matrix multiplication in [[https://​arxiv.org/​abs/​1705.10218 | this manuscript ]]. libDBCSR (or DBCSR for short) is a sparse matrix library designed to efficiently perform sparse matrix matrix multiplication,​ among other operations. It is MPI and OpenMP parallel, and can exploit accelerators. It is developed as part of [[ http://​www.cp2k.org/​ | CP2K ]], where it provides core functionality for [[ doi>​10.1021/​ct200897x | linear scaling electronic structure theory ]]. A general overview of the library [[ doi>​10.1016/​j.parco.2014.03.012 | has been published ]]. A discussion of recent developments,​ in particular GPU work, [[ doi>​10.1002/​9781118670712.ch8 | has appeared as a chapter]] in '​Electronic Structure Calculations on Graphics Processing Units',​ John Wiley and Sons, ISBN 9781118661789,​ and is available as a {{ :​gpu_book_chapter_submitted.pdf |preprint }}. The use of one-sided MPI and a 2.5D algorithm to reduce communication is shown to be effective for sparse matrix matrix multiplication in [[https://​arxiv.org/​abs/​1705.10218 | this manuscript ]].
 +libDBCSR is made available for integration in other projects, see the [[https://​github.com/​cp2k/​dbcsr|github webpage]].
  
-libDBCSR is made available for integration in other projects, ​the latest tarball (nightly update) can be downloaded here:+{{ :​dbcsr_logo.png?​nolink&​300 |}} 
 + 
 +<!--- 
 +the latest tarball (nightly update) can be downloaded here:)
 [[ https://​www.cp2k.org/​static/​dbcsr/​libdbcsr_latest.tgz|{{ :​download-big.png?​direct |Download the latest version}}]] [[ https://​www.cp2k.org/​static/​dbcsr/​libdbcsr_latest.tgz|{{ :​download-big.png?​direct |Download the latest version}}]]
  
Line 12: Line 16:
  
 Developments take place in the [[ download | CP2K svn repository ]], where all version of the library can be found. Developments take place in the [[ download | CP2K svn repository ]], where all version of the library can be found.
 +-->
dbcsr.1516311086.txt.gz · Last modified: 2018/01/18 22:31 by oschuett