Theory and design of portable parallel programs for heterogeneous computing systems and networks

A recurring problem with high-performance computing is that advanced architectures generally achieve only a small fraction of their peak performance on many portions of real applications sets. The Amdahl\u27s law corollary of this is that such architectures often spend most of their time on tasks (codes/algorithms and the data sets upon which they operate) for which they are unsuited. Heterogeneous Computing (HC) is needed in the mid 90\u27s and beyond due to ever increasing super-speed requirements and the number of projects with these requirements. HC is defined as a special form of parallel and distributed computing that performs computations using a single autonomous computer operating in both SIMD and MIMD modes, or using a number of connected autonomous computers. Physical implementation of a heterogeneous network or system is currently possible due to the existing technological advances in networking and supercomputing. Unfortunately, software solutions for heterogeneous computing are still in their infancy. Theoretical models, software tools, and intelligent resource-management schemes need to be developed to support heterogeneous computing efficiently. In this thesis, we present a heterogeneous model of computation which encapsulates all the essential parameters for designing efficient software and hardware for HC. We also study a portable parallel programming tool, called Cluster-M, which implements this model. Furthermore, we study and analyze the hardware and software requirements of HC and show that, Cluster-M satisfies the requirements of HC environments