The goal  is to enable efficient and scalable parallelism.  Current work includes the development of abstractions, algorithms, and language systems.