27 vendors and research organizations now collaborating on developing standard parallel programming model
June 15, 2016 — The University of Bristol has joined the OpenMP ARB, a group of leading hardware and software vendors and research organizations creating the standard for the most popular shared-memory parallel programming model in use today.
“My group is strongly committed to the development and promotion of truly open, cross- platform standards for parallel programming”, said Simon McIntosh-Smith, head of the High Performance Computing group at the University of Bristol. “Developing scientific applications which are capable of exploiting future computer architectures, from multi-core CPUs with wide vector units to many-core GPUs, is already an extremely challenging task. The investment of time and effort required to meet this challenge is best returned by employing standards such as OpenMP, and the early work in my group with partners such as Intel and Cray, has convinced us that OpenMP 4.5 can be a highly effective way of writing portable parallel programs”
The University of Bristol is one of the most popular and successful universities in the UK, and was ranked within the top 40 universities in the world in the QS World University rankings 2015.
“It is good to see that the strong interest of the European OpenMP community also materializes in a growing number of ARB members from the European region”, says Michael Klemm, OpenMP ARB CEO.
The OpenMP ARB now has 13 permanent members and 14 auxiliary members. Permanent members are vendors with a long-term interest in creating products for OpenMP, while auxiliary members have an interest in OpenMP and do not create or sell products.
Call for Papers - OpenMPCon 2016 International OpenMP Developers Conference October 03-05 2016 Nara, Japan http://www.openmpcon.org IMPORTANT DATE: Abstract Deadline: June 11, 2016 ***************************************************************** Sponsored by: OpenMP ARB Description: OpenMPCon is the annual OpenMP Developers conference organized by the OpenMP community, for the community. OpenMPCon takes place 03-05 October in Nara, Japan and will focus on the practical application of OpenMP and other directive based languages, to deliver high-performance computing through parallel processing. Associated workshops, tutorials, special sessions, poster sessions and an exhibition will run in parallel with the conference. Submissions: Submissions are requested from industry experts, developers and academia on all topics related to OpenMP, including but not limited to: application development from any discipline, algorithms, performance analysis, tools and libraries, optimization, verification, portability and education. Sessions can be in the form of technical presentations, tutorials and workshops, posters or panel discussions. All submissions should be in the form of short abstracts (Minimum 300 words - Maximum 1,000 words) that will go through a peer review process. Session materials will be posted on the OpenMPCon website and will be available to conference attendees. Topics of interest include, but are not limited to: * How you Use OpenMP * Guidelines of OpenMP * Tips, tricks, gotchas of OpenMP * Performance of OpenMP * Conformance, build, options of OpenMP compilers/runtime implementations * Latest OpenMP and older releases and how it changes life for users * Accelerator, and heterogeneous computing (GPUs, accelerators, FPGA, etc.) * Tasks, task groups and task dependencies * Scientific, HPC workloads * Commercial, analytic workload * Oil and Gas use of OpenMP * Research on OpenMP (although a good research paper should be submitted to IWOMP) * Manufacturing, games, graphics, operating system, device drivers, and networking domains for OpenMP * Comparisons with other parallel languages * Future of OpenMP * Reduction * Base language (C++, C, Fortran) with OpenMP * Affinity * SIMD and vectors * Memory model, atomics, locks, mutexes, barriers and critical sections in OpenMP * OpenMP APIs and environment variables * Cancellation and Error Model * Technical Reports, TR1, TR2 * Arrays shaping in OpenMP * OpenMP concurrency and parallelism * OpenMP designs * Cancellation * Interoperability * Event programming, futures * Governance, ARB, committee logistics * OpenMP website, discussion groups, forums, twikis * Conferences, language meetings, IWOMP * Speculative execution and transactional memory * Software development tools for OpenMP * Relation with other (MPI, OpenCL, OpenACC, ISO, etc) * History of OpenMP * Tutorial-style sessions
Some new developments regarding OpenMP:
- LLVM Clang is growing towards full OpenMP 4.5 support.
- GCC 6.1 has been released with full support for OpenMP 4.5
- The Parallella Board is now running OpenMP. It has dual-core ARM A9 processors and as co-processor Adapteva Epiphany
- How to do parallel processing with Python and OpenMP
- InsideHPC published an article on OpenMP and OpenCL on Intel Xeon Phi
- An excellent route to invoking C++/OpenMP from R is by using the Rcpp package
GCC 6.1 is a major release containing substantial new functionality not available in GCC 5.x or previous GCC releases. The OpenMP 4.5 specification is fully supported in this new release, and the compiler can be configured for OpenMP offloading to Intel XeonPhi Knights Landing and AMD HSAIL.
See https://gcc.gnu.org/gcc-6/ for details.
OpenMPCon is the annual forum for OpenMP developers to present and discuss the development of real-world applications that leverage the performance of parallel processing through the use of OpenMP. OpenMPCon takes place 3-5 October in Nara, Japan.
Abstract submissions are requested from industry experts, developers, and academia on all topics related to OpenMP, including but not limited to: the experience of developing real-world applications with OpenMP, algorithms, performance analysis, tools and libraries, optimization, verification, portability and education. Sessions can be in the form of short form technical presentations, tutorials, workshops, posters, or panel discussions.
The deadline for the submission of short abstracts is 11 June, 2016 - For additional information visit: http://www.openmpcon.org
OpenMPCon 2016 builds on the success of last year’s inaugural event to provide developers with a unique opportunity to enhance their knowledge and share their practical experience of using the OpenMP API, its supporting tools, and applying the latest advances in parallel programming techniques - said Michael Klemm, OpenMP CEO.
OpenMPCon runs in partnership with IWOMP, the International Workshop on OpenMP. IWOMP focuses on the presentation of unpublished academic research and takes place immediately after OpenMPCon. Both events share the same tutorial sessions on Wednesday 5th October, 2016. For additional information visit: www.iwomp.org
The Board of Directors of the OpenMP Architecture Review Board (ARB) is pleased to announce the appointment of Dr. Michael Klemm as the new CEO of the OpenMP organization, a group of vendors and research organizations creating the standard for one of the most popular shared-memory and embedded parallel programming models in use today. Dr. Klemm succeeds Mr. Michael Wong, who served as the OpenMP CEO for almost 5 years. The Board of Directors and the members of the OpenMP ARB would like to thank Mr. Wong for his dedication and leadership over the past years. During his tenure, the OpenMP language standard added support for coprocessors and accelerators, became more agile by adding a procedure for publishing Technical Reports and yearly ratifications to the standard, and significantly grew in membership.
“We are delighted to welcome Dr. Klemm as our CEO,” says Sanjiv Shah, Chairman of the Board of Directors. “His experience leading many organizational activities as well as several technical features of OpenMP make him well qualified to lead the OpenMP ARB.”
“I look forward to working with Dr. Klemm to ensure the continued success and modernization of OpenMP,” says Dr. Bronis de Supinski, Chair of the OpenMP Language Committee. “Very close collaboration between Bronis de Supinski as the Chair of the OpenMP Language Committee and myself will be key to maintain OpenMP the programming model of choice, not only in HPC but for parallel programming in general,” says Michael Klemm.
Dr. Michael Klemm obtained an M.Sc. in Computer Science in 2003 and received a Doctor of Engineering degree (Dr.-Ing.) from the Friedrich-Alexander-University Erlangen-Nuremberg, Germany, in 2008. Michael Klemm works in the Developer Relations Division at Intel in Germany and his areas of interest include compiler construction, design of programming languages, parallel programming, and performance analysis and tuning. Michael Klemm joined the OpenMP organization in 2009.
The OpenMP ARB announces the twelfth International Workshop on OpenMP (IWOMP), which is the premier forum to present and to discuss issues, trends, recent research ideas and results related to programming with OpenMP. The international workshop affords an opportunity for OpenMP users as well as developers to meet to discuss new ideas and information on this topic.
Submissions of unpublished technical papers detailing innovative, original research and development related to OpenMP are solicited. All topics related to OpenMP are of interest, including OpenMP performance analysis and modeling, OpenMP performance and correctness tools, proposed OpenMP extensions, and OpenMP applications in any domain (e.g., scientific and numerical computation, video games, computer graphics, multimedia, information retrieval, optimization, text processing, data mining, finance, signal and image processing and machine learning).
Advances in technologies, such as multicore processors and OpenMP devices (accelerators such as GPGPUs, DSPs or FPGAs), Multiprocessor Systems on a Chip (MPSoCs), and recent developments in OpenMP itself (e.g., tasking) present new opportunities and challenges for software and hardware developers. Recent advances in the C, C++ and Fortran base languages also offer interesting opportunities and challenges to the OpenMP programming model. IWOMP 2016 particularly solicits submissions in these areas as well as ones that discuss how to apply OpenMP to additional models of parallelism such as event loops.
Recently released Clang 3.8 fully supports OpenMP 3.1 and some features of OpenMP 4.0 and 4.5.
For details, see the Clang 3.8 Release Notes.
Pre-compiled packages can be downloaded from http://llvm.org/releases/download.html#3.8.0