SANTA CLARA, CA, Nov. 7 – The OpenACC standards group today announced increasing momentum behind the OpenACC parallel programming standard by developers, technology providers and new research institutional members. OpenACC provides a quicker and easier path to results for scientists when programming accelerators.

OpenACC is a programming standard for parallel computing using directives, designed to enable millions of scientists around the world to easily take advantage of the transformative power of computing systems equipped with heterogeneous CPU/GPU systems. OpenACC provides the easiest way for scientists, with or without extensive parallel programming expertise, to accelerate their research in a matter of hours using familiar programming models.

“2012 was an outstanding year of growth for OpenACC”, said Duncan Poole, President of the OpenACC standards group. “All of the founding members shipped products compliant to the OpenACC standard, key labs and research organizations signed up as members and developers achieved significant benefit from the portability and performance of the OpenACC directives model.”

At SC12, the members will be demonstrating OpenACC support for accelerators from AMD, Intel, and NVIDIA.

Notable among the recent applications that have benefitted from OpenACC programming:

• Using the CAPS entreprise OpenACC compiler on an NVIDIA Tesla C2050 GPU accelerator, PROMES French Laboratory accelerated by 19X its Heat Transfer Ray-Tracing application in comparison to porting with Sandy Bridge. As a consequence, PROMES is able to efficiently and accurately simulate the entire heat transfer, radiative and convective, in order to design and validate a new prototype.
• Researchers at the University of Bristol are experimenting with parallelizing a Lattice Boltzmann code used for simulating fluid flows. Using OpenACC via the PGI 12.8 compiler, running on a Kepler-based NVIDIA GeForce GTX680 GPU, a 14.5X speedup occurred over scalar OpenMP parallel code running on four cores of an Intel Ivy Bridge CPU. With a similar number of directives as the OpenMP version, ease of porting with the OpenACC is a significant time saver.
• Cray scientists working at Oak Ridge National Laboratory porting the turbulent combustion code S3D achieved outstanding scaling on the recently announced Titan supercomputer. The port consisted of refactoring S3D into a hybrid application consisting of MPI, OpenMP and OpenACC. The port illustrated the benefit of OpenACC directives which represented less than .4% of the refactored code to target the accelerated nodes of Titan.

At SC12, the members will be demonstrating OpenACC support for accelerators from AMD, Intel, and NVIDIA. "Flexibility across multiple platforms while giving options of vendors is one of the key benefits to programming with OpenACC," said Thomas Schulthess, Director of the Swiss National Supercomputing Center. "OpenACC provides support and efficiency for programming heterogeneous HPC systems, including all major architectures used in parallel computing today. OpenACC significantly enhances productivity of those scientists who need to migrate legacy software onto modern computing systems".

During 2012 the founding members--CAPS, Cray, PGI and NVIDIA, were joined by Allinea, Georgia Tech, Oak Ridge National Labs, Rogue Wave and University of Houston. These organizations will help determine the strategy and future direction of the OpenACC standard, including best practices for programming and debugging directives, the use of OpenACC at scale, as well as research intended to improve the robustness and performance of OpenACC compilers.

We are proud to announce our newest members:

• CSCS
• Sandia National Labs

OpenACC will be showcased at SC12 in Salt Lake City by all of the partners. Please visit http://www.openacc.org/sc2012 for an up to date listing of the OpenACC talks and sessions throughout the conference.

About OpenACC

The OpenACC Application Program Interface describes a collection of compiler directives to specify loops and regions of code in standard C and Fortran to be offloaded from a host CPU to an attached accelerator, providing portability across operating systems, host CPUs and accelerators. OpenACC allows programmers to provide simple hints (directives) to the compiler, identifying which areas of code to accelerate. By exposing parallelism to the compiler, directives allow the compiler to do the detailed work of mapping the computation onto the accelerator. OpenACC enables users to create a single code base that runs on heterogeneous many-core accelerators as well as multi-core systems, making scaling application performance easier and more portable than ever. It also offers an ideal way to preserve investment in legacy applications.

-----

Source: OpenACC