As EC++ is intended to be a standard language definition for Parallel C++, it can make no assumptions about the machine architecture. Therefore it is only assumed that the machine architecture on which EC++ can be run is a set of basic computing resources (processors) with associated memory and storage.
These resources can themselves be massively parallel systems, multi processor machines, distributed systems or any other parallel architectures. No assumption can be made that these resources are unique processor/memory pairs, shared memory processors or otherwise.