OpenSHMEM 2018: Fifth Workshop on OpenSHMEM and Related Technologies.

"OpenSHMEM in the era of Extreme Heterogeneity"
August 21-23, 2018

The OpenSHMEM Workshop is an annual event dedicated to the promotion and advancement of the OpenSHMEM programming interface and to help shape its future direction. It is the premier venue to discuss and present the latest developments, implementation technologies, tools, trends, recent research ideas and results related to OpenSHMEM. This year's workshop will explore the ongoing evolution of OpenSHMEM to interoperate with other programming models and heterogeneous architectures. The focus will be on future extensions to improve OpenSHMEM on current and upcoming architectures keeping in mind the trend towards heterogeneity. Although this is an OpenSHMEM specific workshop, we welcome ideas used for other PGAS languages/APIs that may be applicable to OpenSHMEM.
Topics of interest for paper submissions include (but are not limited to) the following areas as they relate specifically to OpenSHMEM:

  • Extensions to support non-volatile, high-bandwidth, and other types of memory
  • Scalable massive multithreading support
  • Experiences with applications from any domain, especially dynamic and irregular applications
  • Extensions to the current specification, including fault-tolerance
  • Hybrid programming models: OpenSHMEM combined with task-based models (e.g. OCR, HPX, ParSEC, etc.) or heterogeneous models (e.g. OpenCL, OpenACC, CUDA, OpenMP, etc.)
  • Low-level communication layers to support OpenSHMEM or other PGAS languages/APIs
  • Big data and I/O support
  • Experiences implementing OpenSHMEM on new architectures
  • Power / Energy studies
  • Static analysis, and verification tools
  • Performance analysis tools for OpenSHMEM and/or other PGAS languages/APIs
  • Auto-tuning or optimization strategies
  • Runtime environments and schedulers
  • Benchmarks, performance evaluation, and validation suites

The workshop proceedings are now available. Please go to to download your copy.