Abstract
Design space exploration is the process of obtaining the optimal design out of all possible design alternatives. The design-space becomes very large when the target architecture consists of heterogeneous components, and it is not possible to explore it in a very short time frame. This thesis attempts to reduce the design-space that has multiplied with the advent of the multiprocessor system-on-chip (MSoC). Traditional methodology suits the single processor system-on-chip, which has fixed allocated components and the embedded-system designer has to partition the specification between hardware and software. However, in MSoC allocation is not fixed. A new stage in the design cycle called "Pre-Allocation" is introduced. It determines the number and type of components required for the implementation at system-level based on performance estimation. Pre-allocation can reduce the design-time for multiprocessor-based embedded systems.
Mohanty, Debashis (2001). Design space exploration for multiprocessor-based embedded systems. Master's thesis, Texas A&M University. Available electronically from
https : / /hdl .handle .net /1969 .1 /ETD -TAMU -2001 -THESIS -M645.