High Performance ParalleX (HPX) is the runtime

implementation of ParalleX, a new model of computation targeting future generation of High Performance Computing Systems. HPX has been designed as an alternative to the conventional computation models such as MPI, attempting to overcome their limitations such as: global barriers, too coarse-grained parallelism and poor latency hiding capabilities. HPX is a modular, feature complete and performance oriented representation of ParalleX. Currently it has been developed on conventional architectures such as Linux or Windows based SMPs or Clusters. The current implementation of HPX provides the infrastructure for key ParalleX concepts: threads and their management, local control objects, parcel transport and parcel management and the active global address. Threads and their Management: The HPX-thread managers implements a work queue based execution model, similar t