Less than six months ago, Wolfram Research brought parallel computing
into the mainstream by integrating automatic parallel technology into
Mathematica 7. Today, that power can be multiplied by more powerful
hardware and clusters with the release of gridMathematica 7

gridMathematica extends the built-in parallelization features of
Mathematica by adding extra computation kernels and automated network
distribution tools — allowing users to run more tasks in parallel,
over more CPUs, for faster execution without changing a line of code,
according to Tom Wickham-Jones, director of Kernel Technology at
Wolfram Research.

gridMathematica Local adds four more computation kernels to the four
already included in a single-machine Mathematica license.
gridMathematica Server provides a network-managed pool of computation
kernels that can be shared by a group of Mathematica users locally or
run on remote hardware to combine the power of multiple computers.

Wolfram Lightweight Grid Manager is designed to make it easy for
users to acquire and run Mathematica computation kernels on remote
hardware and can be used to control high-performance clusters or
create ad hoc grids from idle computers, says Wickham-Jones. Using
gridMathematica 7, users can automatically scale up their tasks to
grids of any size, he adds.

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