Scheduling and data relocation for heterogeneous computing systems and parallel implementations of block-based motion vector estimation for video compression
Abstract
In a dedicated mixed-machine heterogeneous computing (HC) system, an application program may be decomposed into subtasks, then each subtask assigned to the machine where it is best suited for execution. Subtask data relocation is defined as selecting the sources for their needed data items. First, a refinement procedure for data relocation is presented for HC systems in which the communication steps can be modeled as being sequential. Next, a theoretical stochastic HC model is developed, in which the computation times of subtasks and communication times for inter-machine data transfers can be random variables. Given this model, the search space for subtask matching, scheduling, and data relocation is defined and the method for computing execution times of HC applications is derived. A greedy approach to establishing a local optimization criterion for developing data relocation heuristics is validated. Furthermore, the data relocation related concepts and techniques are applied in a data staging problem for scheduling data requests in a heterogeneous information infrastructure. Parallel algorithms, based on a distributed memory machine model, for an exhaustive search technique for block-based motion vector estimation in video compression are designed and evaluated. These algorithms have been implemented on the MasPar MP-1 (a SIMD machine), the Intel Paragon XP/S and IBM SP2 (two MIMD machines), and the PASM prototype (a SIMD/MIMD mixed-mode machine). The trade-offs of using different modes of parallelism and different data partitioning schemes are examined. The analytical and experimental results shown in this application study will help practitioners to predict and contrast the performance of different approaches to parallel implementation of this important video compression technique.
Degree
Ph.D.
Advisors
Siegel, Purdue University.
Subject Area
Electrical engineering|Computer science
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.