This is the archive for November, 2012
Photos from Supercomputing 2012 Salt Lake City (November 2012) are viewable here.
From Supercomputing 2012 Salt Lake City (SC12) (PDFs)
- Progress on the OpenMP Specifications - Bronis de Supinski (November 13, 2012)
- The OpenMP Accelerator Model - James Beyer and Eric Stotzer (November 2012)
- OpenMP on a High Performance DSP - Eric Stotzer (November 2012)
- Fifteen Years of OpenMP - What Next? - Eric Stotzer (November 2012)
OpenMP, the de facto standard for parallel programming on shared memory systems, continues to extend its reach beyond pure HPC to include embedded systems, real time systems, and accelerators.
Release Candidate 1 of the OpenMP 4.0 API specifications currently under development is now available for public discussion. This update includes thread affinity, initial support for Fortran 2003, SIMD constructs to vectorize both serial and parallelized loops,
TASKGROUP, user-defined reductions, and sequentially consistent atomics.
The OpenMP ARB plans to integrate the Technical Report on directives for attached accelerators, as well as more new features, in a final Release Candidate 2, to appear sometime in the first Quarter of 2013, followed by the finalized full 4.0 API specifications soon thereafter.
The 4.0 Release Candidate API specifications (4.0RC1) and the Technical Report (TR1) PDFs can be downloaded from the »OpenMP Specifications page.
A new public »discussion forum has also been created to discuss the 4.0RC1 and the TR1.
Champaign, Illinois - Nov 5, 2012 - The OpenMP Consortium announces the release of a Technical Report detailing directives used for the execution of loops and regions of code on attached accelerators. The directives described in this Technical Report are a work in progress, and the goal of this release is to get early feedback on the proposed directives.
“We aim to provide what the marketplace has been looking for, a standard high-level way of programming accelerators across a broad base of languages and for all forms of accelerator devices”, said Michael Wong, OpenMP CEO.
This Technical Report describes a model for the offloading of code and data onto a target device. Any device may be a target device, including graphics accelerators, attached multiprocessors, co-processors and DSPs. The directives detailed in the Technical Report can be used in Fortran, C, and C++.
The directives are the result of a 3-year effort by the OpenMP Consortium of HPC vendors, national labs, supercomputing centers, academic institutions and users. User experience with members’ initiatives have provided important information to the effort.
The Technical Reports are located at http://openmp.org/wp/openmp-specifications/
Technical Report process
The OpenMP Consortium has created a process by which Technical Reports can be released to the public. With this process, the Consortium will be able to show intermediate versions of the standardization work, and describe possible future directions or extensions to the OpenMP specification. These Technical Reports are not yet part of the OpenMP Standard, in contrast with Specifications, which are normative. Feedback can be posted on the OpenMP Forum, for which registration is required.