Decoupled graph/computation data-driven architecture with variable-resolution actors.

摘要

This paper presents a hybrid multiprocessor architecture that combines the advantages of the dynamic data-flow principles of execution with those of the control-flow model of execution. Two major design ideas are utilized by the proposed model: asynchronous execution of graph and computation operations, and variable- resolution actors. The independence of the two main unites of the machine allows an efficient implementation of functional/data-flow principles with conventional, mature technology. The compiler generates graphs with variable-sized actors in order to match the characteristics of the application to the target machine. For instance, vector actors are proposed for many aspects of scientific computing, while lower resolution (Compound Macro Actors) or conversely higher resolution (atomic instruction actors) is used for unvectorizable programs.