gridMathematica 8 released

Wolfram Research has released gridMathematica 8, the latest version of the company's widely used platform for grid computing.

gridMathematica 8 adds to the existing CPU capabilities, a range of high-performance technologies from CUDA and OpenCL support to dynamically generated C code generation, as well as new application-specific technologies in statistics, image processing, signal processing, finance, and network analysis.

The platform extends the built-in parallelisation features of Mathematica 8 and is able to parallelise and distribute any computation across a grid, making it easy to drive multiple GPU hardware simultaneously across the network from a single control process.

"Mathematica 8's intuitive CUDA programming interface makes it a compelling choice for anyone looking to accelerate high-performance computing applications with GPUs," says Andrew Cresci, General Manager for Strategic Alliances at NVIDIA.

gridMathematica has been at the centre of Wolfram Research's drive to make parallel computing mainstream. With the parallel programming tools built into standard Mathematica licences, over a million people already have access to the tools for submitting tasks to gridMathematica.

"gridMathematica 8 makes further improvements in the automation applied to the distribution of parallel tasks so that developers can focus on the computation, not on details of messaging and synchronisation," says Tom Wickham-Jones, Director of Kernel Technology at Wolfram. "With improvements to the speed of execution of compiled code, including the addition of just-in-time C code generation and compilation, this convenience comes without any performance penalty."

gridMathematica 8 requires Mathematica and is available for Windows, Mac OS X, and Linux.