Uncertainty Model for Configurable Hardware/Software and Resource Partitioning

Rui Wang,William N. N. Hung,Guowu Yang,Xiaoyu Song
DOI: https://doi.org/10.1109/tc.2016.2519895
2016-01-01
Abstract:Automatic hardware/software partitioning relies on characterization, estimation and design space exploration of the system performance and cost metrics. In real world situations, such estimates are complicated and cannot be 100 percent accurate. Furthermore, hardware/software co-design is so complicated nowadays that simply considering the bipartitioning between hardware and software is not sufficient. It is important to consider some of the other key design parameters and resource sharing together with the hardware/software partitioning problem. Under variable requirements of smart systems, more flexibility on the resource usage should be incorporated in system modelling. This paper considers uncertainty modeling for system partitioning with an enhanced set of parameters for hardware/software resource sharing. We harness state-of-the-art uncertainty theory for linear uncertain distribution and normal uncertain distribution. Our derivations convert the uncertainty model back to a regular constraint optimization problem. Experimental results show the effectiveness of our approach.
What problem does this paper attempt to address?