Hardware-Software Partitioning of MPSoC Using Autowave Competition Neural Networks

Sang Nan
2010-01-01
Abstract:A hardware-software partitioning algorithm using autowave competition neural networks is proposed for MPSoCs based on reusable components.The partitioning problem is formulated as a multiconstrained shortest path problem.The autowaves are designed specially to allocate a software component or IP core to each task module from a components library,such that the power consumption of the MPSoC is minimized subject to cost and timing constraints.This algorithm could obtain the globally optimal solution,and it is parallel,non parameter.and easy to implement by VLSI.
What problem does this paper attempt to address?