MULTIOBJECTIVE SYNTHESIS OF LOW-POWER REAL-TIME DISTRIBUTED EMBEDDED SYSTEMS

摘要: This dissertation presents methods for automating the synthesis of embedded systems, i.e., special-purpose computers. In addition, it describes a method analyzing manner in which real-time operating system use influences power consumption. After introducing problem and summarizing previous work field, we present four evolutionary algorithms that simultaneously optimize different costs e.g., price, consumption, response time, area, while ensuring hard constraints are met. These generate multiple solutions tradeoffs between architectural costs. Each algorithm targets domain. The first synthesizes distributed systems. second systems-on-chip composed intellectual property cores may come from vendors. It does clock selection, floorplanning block placement, bus topology generation. third client-server systems bandwidth communication is tightly constrained, wireless incorporates novel scheduling tailored to with clients each server. fourth contain reconfigurable processors. effects usage on overall performance consumption