NVIDIA has released a public beta of the PGI CUDA-enabled Fortran compiler. Developed in collaboration with The Portland Group , it is the first Fortran compiler compatible with NVIDIA CUDA-enabled graphics processing units (GPUs). For more information on CUDA, see the article series CUDA, Supercomputing for the Masses.
GPU computing with the CUDA C-compiler has gained traction in the high-performance computing (HPC) space as it lets developers get transformative increases in performance with minimal coding. However, the case can be made that HPC is still dominated by Fortran, a programming language particularly well suited to numeric computation and scientific computing. Fortran is widely used in applications such as weather modeling, computational fluid dynamics, and seismic processing.
Developers can download the beta CUDA Fortran compiler for Linux, Windows, and Mac OS from The Portland Group website. The PGI CUDA Fortran compiler requires the CUDA SDK which is available from CUDA Zone. For more information on the CUDA Fortran compiler, go here.