CUDA Fortran
PGI CUDA Fortran enables GPU acceleration using the NVIDIA CUDA API and parallel programming model in a native optimizing Fortran 2003 compiler and tool chain.
PGI Accelerator™ C99 & Fortran enable GPU acceleration using OpenMP-like compiler directives in existing Fortran and C99 programs on x64+NVIDIA platforms running Linux, MacOS and Windows.
Multicore Optimizing Parallel Compilers
PGI® Fortran, C and C++ for Linux, MacOS and Windows workstations, servers and clusters based on multicore 64-bit x64, 32-bit x86 and GPU processors.
-
PGI Workstation compilers
and tools suite -
PGI CDK® with cluster-wide
debugging and profiling -
PGI Visual Fortran®
for Microsoft Visual Studio
What's New
- PGI releases PGI Visual Fortran for Visual Studio 2010.
- PGI 2010 version 10.6 is now available. This update includes expanded Fortran 2003 support, PGI Accelerator loop unrolling, PVF support for Visual Studio 2010 and more.
- PGInsider technical newsletter May '10
- PGI compilers now available on Cray CX line of supercomputers.
Optimizing Performance
In the
PGI Accelerator Files you will find articles, tutorials, source code, benchmarks and other useful information to help you with your GPU development efforts.
Use PGI compilers to quickly tune many popular applications, libraries and benchmarks including MPICH & MPICH2, WRF, netCDF, GAMESS, GotoBLAS, OpenMPI, and more.