On Program Restructuring, Scheduling, and Communication for Parallel Processor Systems

220 p.Thesis (Ph.D.)--University of Illinois at Urbana-Champaign, 1986.This dissertation discusses several software and hardware aspects of program execution on large-scale, high-performance parallel processor systems. The issues covered are program restructuring, partitioning, scheduling and interprocessor communication, classification and utilization of program parallelism, and synchronization and hardware design issues of specialized units. All this work was performed focusing on a single goal: maximizing program speedup, or equivalently, minimizing parallel execution time. Parafrase, a Fortran restructuring compiler was used to transform programs in a parallel form and conduct experiments. Two new restructuring techniques are presented, loop coalescing and subscript blocking. Compile-time and run-time scheduling schemes are covered extensively. Depending on the program construct, these algorithms generate optimal or near-optimal schedules. Techniques for program partitioning and minimization of interprocessor communication for idealized program models and for real Fortran programs are also discussed. The close relationship between scheduling, interprocessor communication, and synchronization becomes apparent at several points in this work. Finally, the impact of various types of overhead on program speedup and experimental results are presented.U of I OnlyRestricted to the U of I community idenfinitely during batch ingest of legacy ETD