Computational sprinting: exceeding sustainable power in thermally constrained systems
作者： Milo M. Martin , Arun Raghavan
DOI:
关键词: Transistor 、 Electrical engineering 、 Process (computing) 、 Chip 、 Thermal mass 、 Engineering 、 Exploit 、 Power (physics) 、 Software 、 Mobile device
摘要: Although process technology trends predict that transistor sizes will continue to shrink for a few more generations, voltage scaling has stalled and thus future chips are projected be increasingly power hungry than previous generations. Particularly in mobile devices which severely cooling constrained, it is estimated the peak operation of chip could generate heat ten times faster device can sustainably vent. However, many applications do not demand sustained performance; rather they comprise short bursts computation response sporadic user activity. To improve responsiveness such applications, this dissertation proposes computational sprinting, system greatly exceeds sustainable margins (by up 10×) provide seconds high-performance when interacts with device. Computational sprinting exploits material property thermal capacitance temporarily store excess generated sprinting. After returns levels dissipates stored idle. This dissertation: (i) broadly analyzes thermal, electrical, hardware, software considerations analyze feasibility engineering plat- form 10× higher within today's constraints, (ii) leverages existing sources demonstrate on real today, (iii) identifies energy-performance characteristics determine runtime sprint pacing policies.
参考文章(158)
Evgeni Krimer, Robert Pawlowski, Mattan Erez, Patrick Chiang, Synctium: a Near-Threshold Stream Processor for Energy-Constrained Parallel Applications IEEE Computer Architecture Letters. ,vol. 9, pp. 21- 24 ,(2010) , 10.1109/L-CA.2010.5
Kiran Puttaswamy, Gabriel H. Loh, Thermal Herding: Microarchitecture Techniques for Controlling Hotspots in High-Performance 3D-Integrated Processors high-performance computer architecture. pp. 193- 204 ,(2007) , 10.1109/HPCA.2007.346197
Krisztián Flautner, Steve Reinhardt, Trevor Mudge, Automatic performance setting for dynamic voltage scaling Proceedings of the 7th annual international conference on Mobile computing and networking - MobiCom '01. pp. 260- 271 ,(2001) , 10.1145/381677.381702
Man-Lap Li, R. Sasanka, S.V. Adve, Yen-Kuang Chen, E. Debes, The ALPBench benchmark suite for complex multimedia applications ieee international symposium on workload characterization. pp. 34- 45 ,(2005) , 10.1109/IISWC.2005.1525999
Luiz André Barroso, Urs Hölzle, The Case for Energy-Proportional Computing Computer. ,vol. 40, pp. 33- 37 ,(2007) , 10.1109/MC.2007.443
Daniel Wagner, Dieter Schmalstieg, Making Augmented Reality Practical on Mobile Phones, Part 2 IEEE Computer Graphics and Applications. ,vol. 29, pp. 6- 9 ,(2009) , 10.1109/MCG.2009.67
Francisco Javier Mesa-Martinez, Ehsan K. Ardestani, Jose Renau, Characterizing processor thermal behavior architectural support for programming languages and operating systems. ,vol. 38, pp. 193- 204 ,(2010) , 10.1145/1735971.1736043
Jason Clemons, Andrew Jones, Robert Perricone, Silvio Savarese, Todd Austin, EFFEX: an embedded processor for computer vision based feature extraction design automation conference. pp. 1020- 1025 ,(2011) , 10.1145/2024724.2024949
Abhinav Pathak, Yi-Min Wang, Y. Charlie Hu, Ming Zhang, Paramvir Bahl, Enabling Automatic Offloading of Resource-Intensive Smartphone Applications ,(2011)
Wei Huang, Karthick Rajamani, Mircea R. Stan, Kevin Skadron, Scaling with Design Constraints: Predicting the Future of Big Chips IEEE Micro. ,vol. 31, pp. 16- 29 ,(2011) , 10.1109/MM.2011.42