Filters
Results 1 - 1 of 1
Results 1 - 1 of 1.
Search took: 0.017 seconds
Beyer, J.C.; Stotzer, E.J.; Hart, A.; de Supinski, B.R.
Lawrence Livermore National Laboratory, Livermore, CA (United States). Funding organisation: US Department of Energy (United States)2011
Lawrence Livermore National Laboratory, Livermore, CA (United States). Funding organisation: US Department of Energy (United States)2011
AbstractAbstract
[en] OpenMP (13) is the dominant programming model for shared-memory parallelism in C, C++ and Fortran due to its easy-to-use directive-based style, portability and broad support by compiler vendors. Similar characteristics are needed for a programming model for devices such as GPUs and DSPs that are gaining popularity to accelerate compute-intensive application regions. This paper presents extensions to OpenMP that provide that programming model. Our results demonstrate that a high-level programming model can provide accelerated performance comparable to hand-coded implementations in CUDA.
Primary Subject
Secondary Subject
Source
15 Mar 2011; 16 p; International Workshop on Open MP (IWOMP); Chicago, IL (United States); 13-15 Jun 2011; W-7405-ENG-48; Available from https://e-reports-ext.llnl.gov/pdf/474735.pdf; PURL: https://www.osti.gov/servlets/purl/1022132-IdgFkq/; PDF-FILE: 16; SIZE: 0.2 MBYTES
Record Type
Report
Literature Type
Conference
Report Number
Country of publication
Reference NumberReference Number
INIS VolumeINIS Volume
INIS IssueINIS Issue