Parallel Programming Models

The main goal of the team is to investigate new and current programming paradigms and the associated runtime system support to provide high performance to parallel applications. The target architectures range from multicore and SMT processors to shared- and distributed-memory systems. The Cell BE processor is currently one of our main targets.

• At the programming model, we are proposing extensions to OpenMP to improve the expressiveness of the model. The Mercurium Compiler is able to exploit multiple levels of parallelism and generate work from multiple simultaneously executing threads. Once parallelism is spawned on a coarse level, new opportunities for parallelism on a finer level result in the generation of work for all or for restricted groups of processors. Although our primary focus is on OpenMP, we are also investigating ways to exploit the parallelism in distributed memory architectures, specially now that new systems start offering local memories that must be managed by the software (i.e., in the Cell processor).
• The runtime system offers the basic services to spawn/join parallelism and synchronize across threads. The NthLib API is simple enough to be used for multithreaded programming on a variety of multiprocessor platforms. The goal is to have a platform easy to modify to incorporate new functionalities, as needed by the new architectures. For instance, currently is being used to experiment with new ways to exploit parallelism using local memories in the Cell processor.