Neural Network based On-Chip Thermal Simulator

With increasing power densities, runtime thermal management is becoming a necessity in today's systems, especially so for highly integrated Multi-Processor Systems-on-Chip (MPSoCs). In this paper, the authors propose a Neural Network (NN) based approach to implement an on-chip thermal simulator to aid such runtime management for MPSoCs. The proposed method combines the advantage of approximating the thermal properties of the chip as a linear system with the ease of fully parallel analog implementation of NNs. They perform a case study with the Niagara UltraSPARC T1 MPSoC for real-life applications, benchmarking the results with an accurate higher order Runge-Kutta (RK4) solver, that is employed in tools such as HotSpot.