OpenACC Monthly Highlights: June 2020
Download as PPTX, PDF0 likes822 views
This document provides a monthly highlights summary of OpenACC: - OpenACC is a programming model for parallel computing on CPUs and GPUs using compiler directives to add parallelism to existing serial code. - OpenACC is seeing wide adoption across major HPC applications and allows performance portability between CPU and GPU. - The document highlights recent optimizations, events, publications and resources around OpenACC programming.
OpenACC Monthly Highlights: June 2020
- 2. 2 WHAT IS OPENACC? main() { <serial code> #pragma acc kernels { <parallel code> } } Add Simple Compiler Directive POWERFUL & PORTABLE Directives-based programming model for parallel computing Designed for performance and portability on CPUs and GPUs SIMPLE Open Specification Developed by OpenACC.org Consortium
- 3. 3 silica IFPEN, RMM-DIIS on P100 OPENACC GROWING MOMENTUM Wide Adoption Across Key HPC Codes ANSYS Fluent Gaussian VASP LSDalton MPAS GAMERA GTC XGC ACME FLASH COSMO Numeca 200 APPS* USING OpenACC Prof. Georg Kresse Computational Materials Physics University of Vienna For VASP, OpenACC is the way forward for GPU acceleration. Performance is similar to CUDA, and OpenACC dramatically decreases GPU development and maintenance efforts. We’re excited to collaborate with NVIDIA and PGI as an early adopter of Unified Memory. “ “ VASP Top Quantum Chemistry and Material Science Code * Applications in production and development
- 4. 4 READ BLOG Something that started as a preliminary investigation for an undergraduate class project has become the cover of a prestigious journal publication. Discover how students at the University of Delaware used OpenACC to accelerate molecular dynamics code PPM_One running a large protein complex (approximately 11.3 million atoms) from 14 hours to under 47 seconds. and became the PLOS Computational Biology journal cover. FROM UNDERGRAD PROJECT TO PLOS COVER: OPENACC FOR BIOPHYSICS Image credit: Alex Bryer and Juan R. Perilla
- 5. 5 DON’T MISS THESE UPCOMING EVENTS COMPLETE LIST OF EVENTS Event Call Closes Event Date C-DAC GPU Hackathon July 7, 2020 September 7-11, 2020 Helmholtz GPU Hackathon July 14, 2020 September 14-18, 2020 USTC GPU Bootcamp (Digital) July 21, 2020 September 21-25, 2020 Swiss National Supercomputing Center (Digital) July 12, 2020 September 21-30, 2020 NASA GPU Hackathon July 31, 2020 October 5-10, 2020 NCHC GPU Hackathon (Digital) July 31, 2020 October 12-16, 2020 New in 2020: Many of our events are happening digitally! Get the same high-touch training and mentorship without the hassle of travel!
- 6. 6 REGISTER HPC Summit Digital brings leaders, developers, scientists, and researchers from around the world together to engage with technical experts, ask tough questions, provide feedback, and learn from their peers in an interactive, online setting. Join us to learn about new trends and innovations, engage with experts, and get answers to all your technical questions in our webinars and breakout forums. Best of all, it’s free. LEARN FROM LEADERS IN HPC AT HPC SUMMIT DIGITAL
- 7. 7 READ BLOG The recently concluded GPU Hackathon, hosted by the San Diego Supercomputer Center (SDSC) and held in partnership with Oak Ridge Leadership Computing Facility (OLCF), NVIDIA, and the National Energy Research Scientific Computing Center (NERSC), brought together seven teams across multiple disciplines using a newly launched completely remote, digital event format. By collaborating with mentors who are experts in GPU programming, these seven teams from 11 institutions— many of them having relatively little to no GPU experience—worked to port and optimize their applications. Every team achieved a speed-up. 2020 GPU HACKATHON SERIES KICKS OFF
- 8. 8 LEARN MORE The Programming Models group at Barcelona Supercomputing Center (BSC) has published a new release (version 2020.06) of the OmpSs-2 programming model. The release has several new major features such as a compiler based on LLVM, an integrated tracing tool, and support for OpenACC kernels. Moreover, BSC has optimized the scheduler infrastructure, the memory allocator and the discrete dependency system to improve performance and scalability of OmpSs-2 applications on many-core systems. UPDATE BRINGS NEW FEATURES OMPSS-2 NEW RELEASE
- 9. 9 UPCOMING TALKS FortranCon 2020: Highly Parallel Fortran and OpenACC Directives Jeff Larkin and Michael Wolfe REGISTER Fortran has long been the language of computational math and science and it has outlived many of the computer architectures on which it has been used. Modern Fortran must be able to run on modern, highly parallel, heterogeneous computer architectures. A significant number of Fortran programmers have had success programming for heterogeneous machines by pairing Fortran with the OpenACC language for directives-based parallel programming. This includes some of the most widely-used Fortran applications in the world, such as VASP and Gaussian. This presentation will discuss what makes OpenACC a good fit for Fortran programmers and what the OpenACC language is doing to promote the use of native language parallelism in Fortran, such as do concurrent and Co-arrays.
- 10. 10 UPCOMING TALKS FortranCon 2020: Evolving Fortran for Emerging Architectures: Lessons from the ICON-GPU Atmospheric Model William Sawyer REGISTER For decades Fortran has been on the forefront of high performance computing. As new architectures emerged, the Fortran standard added constructs to exploit them, but not always with complete success. The advent of General Purpose Graphics Processing Units (GPGPUs) has created another conundrum. They can be programmed with an appropriate language (CUDA, CUDAFortran, or OpenCL) or with directives (e.g., OpenMP4.5 or OpenACC3.0), each with disadvantages. In this talk, we outline the lessons learned in porting the ICON atmospheric model to GPUs with OpenCL, CUDAFortran and, finally, OpenACC, with the latter now in production at the Swiss National Supercomputing Centre (CSCS). Now that we understand the programming challenges, it is possible to consider new extensions to the Fortran standard to address GPUs, which are clearly not going away any time soon. We attempt to give some future perspectives.
- 11. 11 RESOURCES Paper: GPU-acceleration of A High Order Finite Difference Code Using Curvilinear Coordinates Marco Kupiainen, Jing Gong, Lilit Axner, Erwin Laure, and Jan Nordström READ PAPER GPU-accelerated computing is becoming a popular technology due to the emergence of techniques such as OpenACC, which makes it easy to port codes in their original form to GPU systems using compiler directives, and thereby speeding up computation times relatively simply. In this study we have developed an OpenACC implementation of the high order finite difference CFD solver ESSENSE for simulating compressible flows. The solver is based on summation-by-part form difference operators, and the boundary and interface conditions are weakly implemented using simultaneous approximation terms. This case study focuses on porting code to GPUs for the most time-consuming parts namely sparse matrix vector multiplications and the evaluations of fluxes. The resulting OpenACC implementation is used to simulate the Taylor-Green vortex which produces a maximum speed-up of 61.3 on a single V100 GPU by compared to serial CPU version.
- 12. 12 RESOURCES Paper: Optimization of Tensor-product Operations in Nekbone on GPUs Martin Karp, Niclas Jansson, Artur Podobas, Philipp Schlatter, and Stefano Markidis In the CFD solver Nek5000, the computation is dominated by the evaluation of small tensor operations. Nekbone is a proxy app for Nek5000 and has previously been ported to GPUs with a mixed OpenACC and CUDA approach. In this work, we continue this effort and optimize the main tensor-product operation in Nekbone further. Our optimization is done in CUDA and uses a different, 2D, thread structure to make the computations layer by layer. This enables us to use loop unrolling as well as utilize registers and shared memory efficiently. Our implementation is then compared on both the Pascal and Volta GPU architectures to previous GPU versions of Nekbone as well as a measured roofline. The results show that our implementation outperforms previous GPU Nekbone implementations by 6-10%. Compared to the measured roofline, we obtain 77 - 92% of the peak performance for both Nvidia P100 and V100 GPUs for inputs with 1024 - 4096 elements and polynomial degree 9. READ PAPER
- 13. 13 STAY IN THE KNOW: JOIN THE OPENACC COMMUNITY JOIN TODAY The OpenACC specification is designed for, and by, users meaning that the OpenACC organization relies on our users’ active participation to shape the specification and to educate the scientific community on its use. Take an active role in influencing the future of both the OpenACC specification and the organization itself by becoming a member of the community.