OpenACC is a user-driven directive-based performance-portable parallel programming model. It is designed for scientists and engineers interested in porting their codes to a wide-variety of heterogeneous HPC hardware platforms and architectures with significantly less programming effort than required with a low-level model. The OpenACC specification supports C, C++, Fortran programming languages and multiple hardware architectures.
ORNL was involved in the formation of the OpenACC organization in order to address the need for a portable programming model to support the OLCF’s first GPU-based supercomputer, Titan. We later became the first non-vendor member of the organization, and continue to be extremely active in guiding the evolution of the OpenACC standard to meet the needs of OLCF users and current and future systems.
Outside of the OLCF, Computer Science and Mathematics Division staff members Seyong Lee and Jeff Vetter have also been an active part of ORNL’s involvement in OpenACC.