Simulation and computer architecture specialist HPC Project has reached the 1.4 version of its Par4All open parallelization platform. The development team's current plan is to extend the platform's scope with regard to the processing and generation of CUDA and OpenCL code.
Par4All is described as an "industrial implementation" of a project developed by teams at France-based CRI/Mines ParisTech and Télécom Bretagne.
NOTE: Par4All is an automatic parallelizing and optimizing compiler (workbench) for C and Fortran sequential programs. The technology's stated goal is to allow industrial users to meet the challenge of multicore architectures and other parallel processors for generic or embedded systems.
The purpose of this source-to-source compiler is to adapt existing applications to various hardware targets such as multicore systems, high-performance computers, and GPUs. It creates a new source code and thus allows the original source code of the application to remain unchanged.
When facing a C or FORTRAN application, Par4All "automatically generates" a portion of parallel code that can be passed to OpenMP, CUDA (which is then suitable for compiler processing on an NVIDIA GPU), and OpenCL. The generated code is then readable and completely traceable with the original code and the whole process works like a typical compiler job.
This new version introduces enhancements for loop processing for CUDA and OpenCL kernel generations. Moreover, dependencies resulting from accesses to global variables are more finely analyzed to assess parallelism.
Underpinned by the group's commercial brand Wild Systems (a mix between hardware and applicative software embedded into a "desktop supercomputing" appliance), the Par4All teams says that it is using its expertise to answer the growing demand for realism in interactive simulation solutions.