To fight against the limits of processors' clock speed, processor manufacturers are now delivering multi-core architectures - with today dual and quad core processors and dozens tomorrow - while most applications, mostly sequential, can only benefit from the power of a single core.

The challenge is now to enable the current and future applications to benefit fully from the new processor technologies by either restructuring existing applications or by facilitating the development of new applications that fully exploit the power of parallelism on these new architectures.

The ParMA project will start by analysing, refining and prioritizing the genuine requirements of the industry. A first version of the ParMA technology will then be introduced at the end of the first project year to gain experience with a wide-range of applications. A second version of ParMA will be released at the end of the second year for further trials.

As parallel programming is becoming mainstream in the era of multi-core architectures, the project achievements will not only be relevant to software engineers who have already developed parallel HPC applications in the past and continue to port and optimize their codes on multi-core architectures: in many domains, parallel code will again be developed and, thus, can largely benefit from this project's achievements. Part of the ParMA software is open source and so available to a wide community.

The three essential goals are to enable:

1. Both the simulation of more complex processes and the modelling of more complex structures without increasing the execution time.

2. SMEs to acquire and run simulation applications that can be hosted on medium-sized systems when the power of multi-core processors is fully exploited.