Data-parallel programming on adaptive and nonuniform computational environments
Date of Award
Doctor of Philosophy (PhD)
Electrical Engineering and Computer Science
adaptive computational environments, parallel processing
In recent years, clusters of machines connected by a high-speed interconnection network are increasingly being used an alternative to more expensive supercomputers for parallelization of data-parallel applications. Usually these clusters of machines are made up of machines with different computational powers. Also, these machines are either dedicated to a single user's computation or shared by users. This thesis describes several software techniques for efficient parallelization of data-parallel applications on adaptive and nonuniform computational environments. It presents several methods for decomposing and mapping structured and unstructured data-parallel applications on static and adaptive nonuniform computational environments. Also, it addresses the runtime support required for the parallelization of such applications. Finally, experimental results for several regular and irregular applications are presented.
Surface provides description only. Full text is available to ProQuest subscribers. Ask your Librarian for assistance.
Kaddoura, Maher, "Data-parallel programming on adaptive and nonuniform computational environments" (1996). Electrical Engineering and Computer Science - Dissertations. 207.