Distributed model predictive control for dynamic thermal management of multi-core systems

Model predictive control (MPC) based dynamic thermal management (DTM) is effective in managing temperature and power of multi-core systems at runtime. However, due to its centralized structure, such method has poor scalability. In this work, we propose a distributed MPC based DTM method, by dividing the full chip thermal model into submodels and decoupling the interactions among submodels with novel techniques. Experiments show that the proposed method performs well in thermal management with fully distributed structure.