SCIENCE

Two important players in the field of parallelism and performance optimization have signed a cooperation agreement. With this agreement, HPC Project will distribute CAPS Entreprise HMPP integrated in a development platform for hybrid computing environments.

The resulting solution includes the Wild Node hardware, a high end configuration of Intel microprocessors and Nvidia GPUs as well as software tools, OS, CAPS HMPP and HPC Project Par4All.

"The complementarity of both our companies approaches will give the developers looking for performance a multi-level solution" says Pierre Fiorini, HPC Project CEO.

For the developer, the solution offers a first level with the Par4All tools for an automated parallelization of a sequential code. From a C or FORTRAN application, Par4All automatically generates a parallel code including CUDA compilable on NVIDIA GPU. The generated code is readable and completely traceable with the original code without any calls to the Par4All tool chain. The original code remains the reference and future other technological choices are open. The whole process works like a usual compilation.

On the second level, HMPP offers an incremental programming model. The developer will be able to take full advantage of hardware accelerators in their native code for a fine tuned performance optimization.

"We are glad to collaborate with HPC Project. We contribute this way to their development integrated solution dedicated to many-core architecture.", says Laurent Bertaux, CAPS CEO.