Join us in Denver for SC17, the International Conference on High Performance Computing where we will be celebrating 20 years of OpenMP as well as releasing the OpenMP 5.0 Beta specification. We also have a packed agenda of other OpenMP events taking place including an OpenMP BOF, booth talks, an exhibitor forum presentation as well as the OpenMP tutorials and talks taking place in the main SC17 program.
- Colorado Convention Center, Denver
- Conference: 12-17 November
- Exhibition: 13-16 November
At the OpenMP Booth: #1246
We are currently finalising this year’s popular Booth Talk series where OpenMP experts share tips, tricks, and the latest on OpenMP.
- hh:mm: Times and Speakers to be confirmed
- hh:mm: Times and Speakers to be confirmed
Join us at 4pm on Tuesday and Wednesday for a beer to celebrate 20 years of OpenMP, hang out with the OpenMP team and hear about the latest developments in the OpenMP API that has been enabling High Performance Computing since 1997.
- Pick up your free reference guides to the OpenMP API (Fortran or C/C++), including OpenMP 4.5.
OpenMP BOF: OpenMP is Twenty. Where is it Going?
- The date, time and location for the BOF is currently awaiting confirmation by the SC17 organisers.
- Session Leaders: Jim Cownie (Intel) and Dr. Michael Klemm (CEO OpenMP ARB / Intel)
- Educate existing and potential users of OpenMP about modern OpenMP and its evolution
- Gather feedback from the audience about where they want OpenMP to go
- Provide a forum for OpenMP users to ask questions of the OpenMP experts who drive the standard
- The BOF will start with a short introduction from Bronis de Supinski, the chair of the OpenMP language committee outlining the major new features planned for OpenMP 5.0. This will be followed by a series of short presentations by other experts from the language committee covering specific topics of interest and will be concluded by Michael Klemm, the OpenMP CEO, who will present the groups vision that drives OpenMP development. The majority of the remaining time (75%) will be open to audience questions to the panel of OpenMP experts.
Papers & Tutorials
- Tutorial: OpenMP Common Core: A Hands-On Exploration
- Sunday, November 12th – 8:30am – 5pm | Location: 303
- Presented by: Timothy Mattson, Alice Koniges, Yun (Helen) He, Barbara Chapman
- Tutorial: Advanced OpenMP: Performance and 4.5 Features
- Sunday, November 12th – 8:30am – 5pm | Location: 404
- Presenters: Christian Terboven, Michael Klemm, Ruud van der Pas, Eric J. Stotzer, Bronis R. de Supinski
- Tutorial: Parallel Computing 101
- Sunday, November 12th – 8:30am – 5pm | Location: 207
- Presented by: Quentin F. Stout, Christiane Jablonowski
- Tutorial: Mastering Tasking with OpenMP
- Monday, November 13th – 8:30am – 12pm | Location: 210-212
- Presented by: Christian Terboven, Michael Klemm, Sergi Mateo Bellido, Xavier Teruel, Bronis R. de Supinski
- Tutorial: Programming Your GPU with OpenMP: A Hands-On Introduction
- Monday, November 13th – 1:30pm – 5pm | Location: 203
- Presented by: Timothy Mattson, Simon McIntosh-Smith
- Paper: An Efficient MPI/OpenMP Parallelization of the Hartree-Fock Method for the Second Generation of Intel Xeon Phi Processor
- Wednesday, November 15th – 4pm – 4:30pm | Location: 402-403-404
- Presented by: Vladimir Mironov, Alexander Moskovsky, Kristopher Keipert, Michael D’mello, Mark Gordon, Yuri Alexeev
- The date, time and location is currently awaiting confirmation by the SC17 organisers.
- Presented by Michael Klemm
- OpenMP: Enabling HPC for Twenty Years
- This presentation reviews the history of OpenMP, the current features and provides an outlook for the upcoming OpenMP API specification, 5.0 Beta, which will be released during SC17. OpenMP is twenty this year. Since its inauguration in 1997, it has become the standard programming model for multi-threading in HPC applications and drove many scientific discoveries by helping domain scientists implement their simulation codes. The OpenMP API is based on directives to augment code written in C/C++ and Fortran with parallelization hints to the compiler. In its 20 years of existence, OpenMP has evolved from a purely multi-threaded model to a programming model that supports modern task-based programming as well heterogeneous programming for offload devices such as GPUs. With it’s extensions it also supports data-parallel programming for single-instruction multiple-data architectures.