Nvidia has released a public beta of the PGI CUDA-enabled Fortran
compile. Developed in collaboration with The Portland Group , it’s
the first Fortran compiler compatible with Nvidia CUDA -enabled
graphics processing units (GPUs).

A compiler is a software tool that translates applications from the
high-level programming languages used by software developers into a
binary form a computer can execute. Developers are invited to
download the beta CUDA Fortran compiler for Mac OS X, Windows and
Linux systems from The Portland Group web site

The PGI CUDA Fortran compiler requires the CUDA Software Developers
Kit (SDK) which is available from CUDA Zone
(http://www.nvidia.com/cuda). PGI compilers are recognized in the HPC
community for delivering world-class performance across a wide
spectrum of applications and benchmarks, according to the folks at