OpenACC Plays a Key Role in the Success of Codes on Summit
The OpenACC organization will report how OpenACC is optimizing CAAR codes are the world’s fastest supercomputer
Dallas, Texas - November 12, 2018 - Today, OpenACC.org announced new milestones featuring OpenACC’s involvement with Oak Ridge National Laboratory's Summit supercomputer. Additionally, the organization announced momentum with HPC apps, the new OpenACC 2.7 specification, new OpenACC.org members and the growth of GPU Hackathons.
OpenACC is a user-driven directive-based performance-portable parallel programming model designed for scientists and engineers interested in porting their codes to a wide variety of heterogeneous HPC hardware platforms and architectures including x86 and OpenPOWER CPUs, GPUs and more with significantly less programming effort.
CAAR Codes Assisting Scientists on Summit
Last June, Oak Ridge National Laboratory announced the Summit Supercomputer, the world’s first supercomputer fusing high performance, data-intensive, and AI computing into one system. Summit is capable of delivering a peak 200 petaflops, eight times faster than its predecessor, Titan.
To take full advantage of Summit, the Oak Ridge Leadership Computing Facility (OLCF) formed the Center for Accelerated Application Readiness (CAAR) program. CAAR is focused on redesigning, porting, and optimizing application codes and several CAAR teams made up of scientists chose to use OpenACC.
“Oak Ridge National Laboratory has been behind OpenACC from its inception. We believe that OpenACC is an important tool to help our users, who are mostly domain scientists, to be able to run their applications on our Titan and Summit supercomputers with excellent performance and significantly less programming effort than required with low-level programming models. Now with Summit, 5 CAAR codes use OpenACC and managed to show a new level of science capability by achieving significant performance results making practical new scientific opportunities.” - Jack Wells, Director of Science for the Oak Ridge Leadership Computing Facility
CAAR codes using OpenACC on Summit include:
- E3SM: Provides high-resolution simulation of the global coupled climate system. “The CAAR project provided us with early access to Summit hardware and access to PGI compiler experts. Both of these were critical to our success. PGI’s OpenACC support remains the best available and is competitive with much more intrusive programming model approaches.” - Mark A. Taylor, Multiphysics Applications, Sandia National Laboratory
- LSDalton: A linear-scaling Hartree-Fock (HF) and Density-Functional Theory (DFT) code suitable for large molecular systems. It provides a 3-5x speedup versus the equivalent number of Titan nodes. “Using OpenACC, we see large performance gains with very little effort. GPU acceleration varies over the course of our simulations due to differing fragment sizes, but is typically 3x–5x. On Summit we can now do simulations of several thousand atoms, compared to maybe 800 on Titan.” - Dmytro Bykov, Computational Scientist, Oak Ridge National Laboratory
- FLASH: A component-based, massively parallel, adaptive-mesh refinement code. There is a 2.9x speedup on GPUs versus CPU only systems. “We’re using OpenACC on Summit to accelerate our most compute-intensive kernels. We love OpenACC interoperability and how this allows us to use multiple methods to perform memory placement and movement. CPU+GPU performance of a 288 species network on Summit, something impossible to do on Titan, is 2.9x faster than CPU only.” - Bronson Messer, Senior Scientist, Oak Ridge National Laboratory
- GTC: A code enabling particle turbulence simulations for sustainable fusion reactions in ITER. “Using OpenACC, our scientists were able to achieve the acceleration needed for integrated fusion simulation with a minimum investment of time and effort in learning to program GPUs.” - Zhihong Lin, Professor and Principal Investigator, UC Irvine
- XGC: A code enabling multiphysics magnetic fusion reactor simulator, from hot core to cold wall. It provides an 11x speedup on GPUs versus CPU only systems. “Using a combination of CUDA and OpenACC for our most compute-intensive kernels, the GPU-accelerated version of XGC delivers over 11x speed-ups compared to CPU-only execution when running at scale on 2048 nodes of ORNL’s new Summit supercomputer.” - C-S Chang, Principal Investigator, Princeton Plasma Physics Lab, Princeton University.
Cream of the Crop HPC Applications Adopting OpenACC
There has been a sustained growth of HPC applications adopting OpenACC because of its relative ease of implementation. Beyond performance improvements, major benefits cited by both researchers and commercial software developers are its portability across architectures and its relative ease of use. OpenACC is now supporting over 150 codes including three of the top five HPC applications. Notable commercial applications include:
- NUMECA FINE/Open: Computational fluid dynamics applications used to model supersonic compressors for large-scale Carbon Capture, Utilization and Sequestration
- Synopsys: Modeling software for camera image sensors
- Gaussian: The most widely used quantum chemistry application for modeling the electronic structure of atoms and molecules
- ANSYS Fluent (in development): Discrete Ordinate Radiation Model
- ADS CFD: Computational fluid dynamics application used in aerospace to model airflow over objects.
The Launch of OpenACC 2.7 Specification
OpenACC 2.7 specification has just been approved. User requested features have been added including ‘self’ clause on compute construct and associated explanation. View the full specification here.
Additionally, PGI 18.10 Community Edition is also now available. New OpenACC features include beta support for true deep Copy, CPU/GPU auto-compare and more. See the PGI new features page for a complete list.
Three New OpenACC Members
SUSE, C-DAC and Osaka University are the three most recent institutions to join the OpenACC membership. The new members will contribute to technical and marketing committees, shape the OpenACC specification to support their research and will help grow a community of OpenACC users who aim to perform more science and research and less programming effort.
GPU Hackathons which started as OpenACC-only events under Oak Ridge National Laboratory umbrella have now grown into a series of events with 160 teams participating from all around the world. A majority of the teams choose OpenACC to start programming GPU, but any GPU programming models and tools are welcome at the events.
This month, Oak Ridge National Laboratory is announcing their first 10 hackathons in 2019. The event held in KISTI Supercomputing center opened its call for applications today: https://www.olcf.ornl.gov/gpu-hackathons/.
ORNL will be adding more events to the calendar in the coming months.
About the OpenACC.org
OpenACC.org is a nonprofit organization founded to help scientists and researchers do more science and less programming by providing a high-level directives-based programming model for high-performance computing. The charter of OpenACC.org is to develop and deliver the OpenACC specification and to help educate and support scientists and researchers using or considering using OpenACC directives. The organization is run by academic and industry members in collaboration with the OpenACC User Community. Please visit www.openacc.org for more information.
OpenACC PR Director