The Portland Group (PGI), a wholly-owned subsidiary of STMicroelectronics, has announced the general availability of PGI Visual Fortran (PVF) for Visual Studio 2010. PVF integrates PGI high-performance parallel Fortran compilers and tools with Microsoft Visual Studio to offer a high-productivity development solution to scientists and engineers upgrading to the latest 64-bit multi-core platforms running Microsoft Windows.
PGI compilers and tools are used widely by performance-oriented programmers on Linux, Mac OS X and Windows systems based on multi-core CPUs from Intel and AMD and incorporating GPU accelerators from Nvidia. The new 10.6 version of the PGI 2010 release adds support for building Windows Fortran applications using the latest version of the popular Microsoft Visual Studio integrated development environment (IDE) - Visual Studio 2010. Visual Studio is the most widely used IDE in the world. PVF tools and technologies, including an MPI/OpenMP parallel debugger, enable Visual Studio developers to efficiently develop High Performance Computing (HPC) applications for multi-core workstations and Windows HPC Server 2008 clusters. In addition, PGI Visual Fortran is available with support for programming Nvidia GPU accelerators using directive-based PGI Accelerator Fortran or CUDA Fortran language extensions.
Key features of the new compilers include: Native Fortran 95/2003 Visual Studio project system; Fortran text editor extensions, intrinsic function tips and keyword completion; integrated PGI-custom Fortran-capable debugger for debugging of single-thread, multi-thread and OpenMP parallel applications and MSMPI parallel applications running locally and on clusters; Microsoft Visual C++ interoperability and full support for Visual Studio 2010, Visual Studio 2008 and Visual Studio 2005; and optional support for the Visual Numerics IMSL Fortran Numerical Library and the Intel MKL Math Kernel Library.
In addition to comprehensive support for multi-core CPUs, PGI Visual Fortran also supports GPU accelerators. The PGI Accelerator programming model is a high-level implicit model similar to OpenMP for multi-core x64 systems. Using compiler directives, programmers can offload compute-intensive code regions from a host CPU to a GPU accelerator. Programs containing PGI Accelerator directives remain 100% standard-compliant and portable. PVF also supports CUDA Fortran language extensions. Defined in cooperation with Nvidia, CUDA Fortran is an analog to the Nvidia CUDA C compiler and gives expert programmers direct control of all aspects of GPGPU programming.