Abstract

This thesis focuses on the design and implementation of Concordia Parallel C (CPC) and its compiler Concordia Parallel C Compiler (CPCC). The Concordia Parallel Programming Environment (CPPE) is a novel parallel programing environment supporting virtual-architecture parallel programming paradigm, program and library development, simulated execution, syntax/semantics/performance debugging, and simulation of various multiprocessors and multicomputers. A Parallel program written in CPC will be compiled by CPCC into a virtual code version, then interpreted by Concordia Parallel Systems Simulator (CPSS). We extend the C language with features supporting Parallel computing as well as selected features of C++ to facilitate Parallel library writing. We adopt Abstract Syntax Trees (AST) as the internal code representation of our CPCC front-end, from which various program transformations can be performed, and code for various target parallel systems can be generated. The current back-end generates virtual code for CPSS.

Document Downloads