Users of the Linux GCC compiler will soon have access to GPU accelerator programming thanks to the work of the OpenACC Standards Group and Mentor Graphics.
Mentor and the Oak Ridge Leadership Computing Facility are collaborating to integrate the OpenACC standard into the open-source GCC compiler suite. OpenACC is an application program interface that allows programmers to provide directives to the compiler to identify what code to offload to the GPU accelerator. The GCC compiler was produced by the GNU Project.
According to Mentor, since GCC is the default compiler on most Linux distributions, the integration could expand access to the language and “facilitate the development and testing of OpenACC applications on smaller systems, such as workstations and clusters.”
The extensions for GCC came via Mentor’s Code Sourcery acquisition. Mentor is currently working on implementing OpenACC 2.0 in the GCC toolchain for Linux users. The company and the OpenACC group hope this can help move accelerator use out of the science niche and into additional applications and markets.
“The addition of an open-source platform is critical for the broader adoption of heterogeneous programming that will be necessary for the next stage of growth and innovation in HPC and as a foundation to drive software development for Exascale systems,” said Duncan Poole, president of OpenACC.org, in an announcement and slidecast last week.
Mentor just recently joined the OpenACC organization, along with Louisiana State University and the National Oceanic and Atmospheric Administration (NOAA). OpenACC has been used on the Titan supercomputer at Oak Ridge to accelerate a number of applications to GPUs.
Source: HPC Wire