Compile a graph-based program specification with automated clustering of graph components based on identifying specific data port connections

摘要

A graph-based program specification (628) is a plurality of components, each corresponding to a processing task, including one or more ports, for the port to send or receive a single data element A plurality of components, including a scalar data port and a collection data port for sending or receiving a collection of data elements, each connecting an upstream component output port to a downstream component input port Contains one or more links. Prepared code representing a plurality of component subsets (630, 631) is generated to identify one or more links connecting the component collection data port to the component scalar data port. Including one or more subset boundaries, forming a subset based on the identified subset boundaries, and each formed subset when used for execution by a runtime system Generating prepared code for each formed subset that causes the processing of tasks corresponding to the components within to be performed.