2010 ACM/IEEE International Conference for High Performance Computing, Networking, Storage and Analysis 2010
DOI: 10.1109/sc.2010.40
|View full text |Cite
|
Sign up to set email alerts
|

Parallelizing the QUDA Library for Multi-GPU Calculations in Lattice Quantum Chromodynamics

Abstract: Graphics Processing Units (GPUs) are having a transformational effect on numerical lattice quantum chromodynamics (LQCD) calculations of importance in nuclear and particle physics. The QUDA library provides a package of mixed precision sparse matrix linear solvers for LQCD applications, supporting single GPUs based on NVIDIA's Compute Unified Device Architecture (CUDA). This library, interfaced to the QDP++/Chroma framework for LQCD calculations, is currently in production use on the "9g" cluster at the Jeffer… Show more

Help me understand this report
View preprint versions

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1
1
1

Citation Types

1
82
0

Year Published

2012
2012
2022
2022

Publication Types

Select...
7
2
1

Relationship

0
10

Authors

Journals

citations
Cited by 88 publications
(85 citation statements)
references
References 26 publications
(44 reference statements)
1
82
0
Order By: Relevance
“…The software codes Chroma [58] and QUDA [59,60] were used to perform this work at Jefferson Laboratory under the USQCD Initiative and the LQCD ARRA project, and on the Lonsdale cluster maintained by the Trinity Centre for High Performance Computing funded through grants from Science Foundation Ireland (SFI). This work also used the DiRAC Data Analytic system at the University of Cambridge, operated by the University of Cambridge High Performance Computing Service on behalf of the STFC DiRAC HPC Facility (www.dirac.ac.uk).…”
Section: Discussionmentioning
confidence: 99%
“…The software codes Chroma [58] and QUDA [59,60] were used to perform this work at Jefferson Laboratory under the USQCD Initiative and the LQCD ARRA project, and on the Lonsdale cluster maintained by the Trinity Centre for High Performance Computing funded through grants from Science Foundation Ireland (SFI). This work also used the DiRAC Data Analytic system at the University of Cambridge, operated by the University of Cambridge High Performance Computing Service on behalf of the STFC DiRAC HPC Facility (www.dirac.ac.uk).…”
Section: Discussionmentioning
confidence: 99%
“…Systems may be considered having non-zero overall momentum, P = 2π L n x , n y , n z , which satisfy periodic boundary conditions if n x , n y , n z are integers. Such 1 The large number of propagators are very efficiently computed using an Adaptive Multi-Grid solver [32,33] for the light quarks on CPUs, and the strange quark propagators are efficiently computed using Graphical Processing Units [34,35]. Table I.…”
Section: Calculating the Finite-volume Spectrummentioning
confidence: 99%
“…Support for GPUs is based on the QUDA library [3], originally developed at Boston University (BU). The extension to staggered quarks began when one of us was on sabbatical at the National Center for Supercomputing Applications (NCSA) at the University of Illinois.…”
Section: Development Of the Milc Codementioning
confidence: 99%