Aug. 3 — Directive based programming models offer scientific applications a path onto HPC platforms without undue loss of portability or programmer productivity. Using directives, application developers can port their codes to the accelerators incrementally while minimizing code changes. Challenges remain because the directives models need to support a rapidly evolving array of hardware with diverse memory subsystems, which may or may not be unitied. The programming model will need to adapt to such developments, make improvements to raise its performance portability that will make accelerators as first-class systems for HPC. Such improvements are being continuously discussed with the standards committees such as OpenMP and OpenACC. This workshop airs to capture the assessment of the improved feature set, their implementation and experiences with their deployment in HPC applications. The workshop aims at bringing together the user and tools community to share their knowledge and experiences of using directives to program accelerators,
An NVIDIA Quadro M6000 will be awarded to the principal author of the Best Paper submitted, as determined by the Program Chairs.
The OpenACC Toolkit, including a free PGI Accelerator Compiler for academics, is available for those who need programming tools.
Submission Guide
Format: Submissions are limited to 10 pages in the ACM format (see http://www.acm.org/sigs/publications/proceedings-templates).
The 10-page limit includes figures, tables, and appendices, but does not include references, for which there is no page limit.
Submission Topics
Topics of interest for workshop submissions include (but are not limited to):
- Experience porting applications in any domain using directives
- Extensions to and shortcomings of current accelerator directives APIs
- Hybrid heterogeneous or many-core programming with accelerator directives with other models (OpenMP, MPI, OpenSHMEM)
- Scientific library interoperability with accelerator directives
- Experiences in implementating compilers for accelerator directives on new architectures
- Low level communications APIs or runtimes that support accelerator directives
- Asynchronous execution and scheduling (heterogeneous tasks)
- Extensions to programming models supporting memory heirarchies
- Performance evaluation
- Power/energy studies
- Static analysis and verification tools
- Modeling and performance analysis tools
- Auto-tuning or optimization strategies
- Benchmarks and validation suites
Papers Submission Guidelines:
We plan to use Easychair submission site for paper submission. Link: https://easychair.org/conferences/?conf=waccpd15
Important Dates:
- Extended to: August 28th, 2015 (Midnight 11:59 Pacific Time Zone)
- Author notification: September 30th, 2015
- Camera Ready papers due: October 7th, 2015
Keynote Speaker:
Barbara Chapman, a Professor in the Department of Computer Science and Director of the Center for Advanced Computing and Data Systems (CACDS), University of Houston, Houston, TX, USA has been engaged in research on parallel programming languages and compiler technology for more than 15 years. Her research group has developed OpenUH, a state-of-the art open source compiler that is used to explore language, compiler and runtime techniques, with a special focus on multithreaded programming. Dr. Chapman has been involved with the evolution of the OpenMP directive-based programming standard since 2001. She also is a participant in the OpenSHMEM and OpenACC programming standards efforts. Her on-going research continues to advance these efforts. It also explores new approaches to optimize partitioned global address space programs, strategies for runtime code optimizations, compiler-tools interactions and high-level programming models for embedded systems.
Prof. Chapman completed her Ph.D. on software support for distributed memory programming at Queen’s University of Belfast, Northern Ireland, U.K.
More information can be found here: http://www.openacc.org/content/Events/waccpd_2015
—
Source: OpenACC