D3. FPGA Implementation of 3-D Fuzzy Logic Controller for Multi-core CPU Thermal Management

Girgis, M. E., M. B. Abdelhalim, and H. A. Kamal, "D3. FPGA Implementation of 3-D Fuzzy Logic Controller for Multi-core CPU Thermal Management", 2013 30th National Radio Science Conference (NRSC), Egypt, pp. 448-457, April, 2013. copy at www.tinyurl.com/yag2677g

Date Presented:



A multi-core processor is an Integrated Circuit (IC) which two or more processors have been attached inside a single package for enhanced performance, reduced power consumption, and more efficient simultaneous processing of multiple tasks. [n fact, the chip power consumption is limited by cooling system level capacity. The air cooling limitation is already reached in 2006. The CPU reaches the maximum operational temperature after certain time due to maximum CPU utilization, thus the CPU utilization is reduced to the safe utilization in order not to exceed the power limit, and this phenomenon is called CPU thermal throttling. The cores thermal problem is managed by adding temperature constraints through a threshold temperature (maximum temperature). This problem is related to the power consumption management that is a function of the operating frequency of each core where damaging one core means damaging the whole chip. The current software solutions of thermal management problems are not effective as they are slow and don' t take into the consideration the mutual heat transfer between the cores i.e. the correlation. Therefore, a hardware solution is proposed in this paper that takes into consideration the mutual heat transfer between the cores. This solution is built using the three dimensional fuzzy logic controller. The obtained results prove that the 3-D fuzzy logic controller is able to process the multicore CPU correlation information to improve the thermal management response and reduce the air cooling limitation effect. The controller is implemented using FPGAs technology where different architectures are explored according to the given constraints.



Related External Link