DocumentCode :
2649838
Title :
Architecture level thermal modeling for multi-core systems using subspace system method
Author :
Eguia, Thorn Jefferson A ; Shen, Riujing ; Tan, Sheldon X D ; Pacheco, Eduardo H. ; Tirumala, Murti
Author_Institution :
Dept. of Electr. Eng., Univ. of California, Riverside, CA, USA
fYear :
2009
fDate :
20-23 Oct. 2009
Firstpage :
714
Lastpage :
717
Abstract :
This paper proposes a new architecture-level thermal modeling method to address the emerging thermal related analysis and optimization problem for high-performance multi-core microprocessor design. The new approach, called Thermsid, builds the thermal behavioral models from the measured or simulated thermal and power information at the architecture level. Thermsid builds the thermal models by first generating a Hankel matrix of Markov parameters, from which state matrices are obtained through minimum square optimization. Compared to existing behavioral thermal modeling algorithms, the proposed method is much more general as it does not require step temperature responses, leading to greater flexibility during the modeling process. Experimental results on a real quad-core microprocessor show that Thermsid provides accurate thermal behavioral models comparable existing approaches.
Keywords :
Hankel matrices; circuit optimisation; integrated circuit modelling; microprocessor chips; thermal analysis; Hankel matrix; Markov parameters; Thermsid; architecture level thermal modeling; high-performance multi-core microprocessor design; minimum square optimization; quad-core microprocessor; state matrices; subspace system method; thermal analysis; Design optimization; Microprocessors; Multicore processing; Packaging; Power measurement; Power system modeling; Rapid thermal processing; System identification; Temperature measurement; Thermal conductivity; Thermal analysis; architecture thermal modeling; multicore processor;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
ASIC, 2009. ASICON '09. IEEE 8th International Conference on
Conference_Location :
Changsha, Hunan
Print_ISBN :
978-1-4244-3868-6
Electronic_ISBN :
978-1-4244-3870-9
Type :
conf
DOI :
10.1109/ASICON.2009.5351305
Filename :
5351305
Link To Document :
بازگشت