Date: 23/06/2009

ST's PGI and NVIDIA join to deliver CUDA Fortran compiler

ST Micro's subsidiary Portland Group has signed agreement with NVIDIA under which the two companies plan to develop new Fortran language support for CUDA GPUs.

The NVIDIA CUDA architecture parallelize the kernels utilizing the processing power of Graphic Processing Unit (GPU). CUDA offers developers explicit control over the mapping of general-purpose computational kernels to GPUs and placement and movement of data between the x64 processor and the GPU. The CUDA Fortran compiler will provide the same level of control and optimization in a native Fortran environment from PGI as it is provided in NVIDIA CUDA C compiler.

"Fortran support for CUDA GPUs is a perfect complement to our existing roadmap for the PGI Accelerator Fortran and C compilers," said Douglas Miles, director, The Portland Group. "It enables interoperability of PGI Fortran and CUDA C and gives PGI users a full range of options in porting and optimizing Fortran applications to leverage the power of CUDA-enabled NVIDIA GPUs."

"The GPU computing developer community has made it clear there is a need and demand for a production-quality Fortran solution on the GPU," said Andy Keane, general manager, Tesla GPU Computing Solutions, NVIDIA. "With their large base of Fortran developers for x64 processor-based HPC systems, PGI provides a perfect bridge for migration of production science and engineering codes from existing platforms to NVIDIA Tesla GPUs."

The Portland Group and NVIDIA will release the Fortran language specification for CUDA GPUs at the International Conference on Supercomputing in Hamburg, Germany this week. The CUDA Fortran compiler will be added to a production release of the PGI Fortran compilers scheduled for availability in November 2009. More detailed information about PGI compilers and tools is available online at http://www.pgroup.com.

The use of Graphics Processing Units (GPUs) as general purpose accelerators has been a growing trend in high-performance computing (HPC). Until now, use of GPUs from Fortran applications has been extremely limited. Developers targeting GPU accelerators have had to program in C at a detailed level using sequences of function calls to manage movement of data between the x64 host and GPU, and to offload computations from the host to the GPU. The PGI Accelerator Fortran and C compilers automatically analyze whole program structure and data, split portions of an application between a multi-core x64 CPU and a GPU as specified by user directives, and define and generate a mapping of loops to automatically use the parallel cores, hardware threading capabilities and SIMD vector capabilities of modern GPUs.


The PGI has also announced the availability of the PGI Release 9.0 line of high-performance parallelizing compilers and development tools for Linux, Mac OS X and Windows. PGI Release 9.0 is the first general release to include support for the high-level PGI Accelerator programming model on x64 processor-based Linux systems incorporating NVIDIA CUDA-enabled GPUs.