ABSTRACT
Microprocessor voltage levels include substantial margin to deal with process variation, system power supply variation, workload induced thermal and voltage variation, aging, random uncertainty, and test inaccuracy. This margin allows the microprocessor to operate correctly during worst-case conditions, but during typical conditions it is larger than necessary and wastes energy. We present a mechanism that reduces excess voltage margin by (1) introducing a critical path monitor (CPM) circuit that measures available timing margin in real-time, (2) coupling the CPM output to the clock generation circuit to adjust clock frequency within cycles in response to excess or inadequate timing margin, and (3) adjusting the processor voltage level periodically in firmware to achieve a specified average clock frequency target. We implemented this mechanism in a prototype IBM POWER7 server. During better-than-worst case conditions our guardband management mechanism reduces the average voltage setting 137-152 mV below nominal, resulting in average processor power reduction of 24% with no performance loss while running industry-standard benchmarks.
- Bowman, K. A.; Tschanz, J. W.; Kim, N. S.; Lee, J. C.; Wilkerson, C. B.; Lu, S.-L.L.; Karnik, T.; De, V. K. 2009. Energy-Efficient and Metastability-Immune Resilient Circuits for Dynamic Variation Tolerance. IEEE Journal of Solid-State Circuits. 44, 1 (Jan. 2009), 49--63. DOI=10.1109/JSSC.2008.2007148.Google ScholarCross Ref
- Das, S.; Tokunaga, C.; Pant, S.; Ma, W.-H.; Kalaiselvan, S.; Lai, K.; Bull, D. M.; Blaauw, D. T. 2009. RazorII: In Situ Error Detection and Correction for PVT and SER Tolerance. IEEE Journal of Solid-State Circuits, 44, 1 (Jan. 2009). IEEE, 32--48. DOI=10.1109/JSSC.2008.2007145.Google ScholarCross Ref
- Das, S. 2009. Razor: A Variability-Tolerant Design Methodology for Low-power and Robust Computing. Doctoral Thesis. University of Michigan. Google ScholarDigital Library
- Dighe, S.; Vangal, S.; Aseron, P.; Kumar, S.; Jacob, T.; Bowman, K.; Howard, J.; Tschanz, J.; Erraguntla, V.; Borkar, N.; De, V.; Borkar, S. 2010. Within-die variation-aware dynamic-voltage-frequency scaling core mapping and thread hopping for an 80-core processor. International Solid-State Circuits Conference Digest of Technical Papers (Feb. 7--11, 2010). ISSCC 2010. IEEE, 174--175. DOI=10.1109/ISSCC.2010.5433997.Google ScholarCross Ref
- Drake, A.; Senger, R.; Deogun, H.; Carpenter, G.; Ghiasi, S.; Nguyen, T.; James, N.; Floyd, M.; Pokala, V. 2007. A Distributed Critical-Path Timing Monitor for a 65nm High-Performance Microprocessor. International Solid-State Circuits Conference Digest of Technical Papers (Feb 11--15, 2007). ISSCC 2007. IEEE, 398--399. DOI=10.1109/ISSCC.2007.373462.Google ScholarCross Ref
- Ernst, D.; Kim, N. S.; Das, S.; Pant, S.; Rao, R.; Pham, T.; Ziesler, C.; Blaauw, D.; Austin, T.; Flautner, K.; Mudge, T. 2003. Razor: a low-power pipeline based on circuit-level timing speculation. In Proceedings of the 36th Annual IEEE/ACM International Symposium on Microarchitecture (Dec 3--5, 2003). MICRO-36. 7--18. DOI=10.1109/MICRO.2003.1253179. Google ScholarDigital Library
- Ernst, D.; Das, S.; Lee, S.; Blaauw, D.; Austin, T.; Mudge, T.; Kim, N. S.; Flautner, K. 2004. Razor: circuit-level correction of timing errors for low-power operation. Micro, 24, 6 (Nov.-Dec. 2004), IEEE, 10--20. DOI=10.1109/MM.2004.85. Google ScholarDigital Library
- Fick, D.; Liu, N.; Foo, Z.; Fojtik, M.; Seo, J.; Sylvester, D.; Blaauw, D. 2010. In situ delay-slack monitor for high-performance processors using an all-digital self-calibrating 5ps resolution time-to-digital converter. International Solid-State Circuits Conference Digest of Technical Papers (Feb. 7--11, 2010). ISSCC 2010. IEEE, 188--189. DOI=10.1109/ISSCC.2010.5433996.Google ScholarCross Ref
- Fischer, T.; Anderson, F.; Patella, B.; Naffziger, S. 2005. A 90nm variable-frequency clock system for a power-managed Itanium®-family processor. International Solid-State Circuits Conference Digest of Technical Papers (Feb. 2005). ISSCC 2005. IEEE. 294--299. DOI=10.1109/ISSCC.2005.1493985.Google ScholarCross Ref
- Floyd, M.; Allen-Ware, M.; Rajamani, K.; Brock, B.; Lefurgy, C.; Drake, A. J.; Pesantez, L.; Gloekler, T.; Tierno, J. A.; Bose, P.; Buyuktosunoglu, A. 2011. Introducing the Adaptive Energy Management Features of the Power7 Chip. IEEE Micro, 31, 2, (March-April 2011), 60--75. DOI=10.1109/MM.2011.29. Google ScholarDigital Library
- Gupta, M. S.; Rangan, K. K.; Smith, M. D.; Gu-Yeon Wei; Brooks, D. 2008. DeCoR: A Delayed Commit and Rollback mechanism for handling inductive noise in processors. In IEEE 14th International Symposium on High Performance Computer Architecture (Salt Lake City, UT, February 2008). HPCA 2008. IEEE, 381--392. DOI=10.1109/HPCA.2008.4658654.Google ScholarCross Ref
- IBM. 2011. IBM POWER7 Technology and Systems. IBM Journal of Research and Development. 55, 3 (May-June 2011). DOI=10.1147/JRD.2011.2128750.Google Scholar
- Intel Corporation. 2009. Voltage Regulator Module (VRM) and Enterprise Voltage Regulator-Down (EVRD) 11.1 Design Guidelines. Reference Number 321736, Revision 002, September, 2009.Google Scholar
- Isci, C.; Contreras, G.; Martonosi, M. 2006. Live, Runtime Phase Monitoring and Prediction on Real Systems with Application to Dynamic Power Management. In Proceedings of the 2006 39th Annual IEEE/ACM International Symposium on Microarchitecture (MICRO 39). IEEE Computer Society, Washington, DC, USA, 359--370. DOI=10.1109/MICRO.2006.30. Google ScholarDigital Library
- Kang, I.; Ethirajan, K.; Severson, M. 2005. Dynamic Voltage Scaling for Portable Devices, U. S. Patent Application 2005/0218871 A1.Google Scholar
- Keane, J.; Kim, C. H. 2001. An odometer for CPUs. IEEE Spectrum. 48, 5 (May 2011), 28--33. DOI=10.1109/MSPEC.2011.5753241.Google Scholar
- Meterelliyoz, M.; Goel, A.; Kulkarni, J. P.; Roy, K. 2010. Accurate characterization of random process variations using a robust low-voltage high-sensitivity sensor featuring replicabias circuit. International Solid-State Circuits Conference Digest of Technical Papers (Feb. 7--11, 2010). ISSCC 2010. IEEE, 186--187. DOI=10.1109/ISSCC.2010.5433991.Google Scholar
- Reddi, V.; Kanev, S.; Kim, W.; Campanoni, S.; Smith, M. D.; Wei, G.-Y.; Brooks, D. 2010. Voltage Smoothing: Characterizing and Mitigating Voltage Noise in Production Processors via Software-Guided Thread Scheduling. In Proceedings of the 2010 43rd Annual IEEE/ACM International Symposium on Microarchitecture (MICRO 43). IEEE Computer Society, Washington, DC, USA, 77--88. DOI=10.1109/MICRO.2010.35. Google ScholarDigital Library
- Rylyakov, A.; Tierno, J.; English, G.; Sperling, M.; Friedman, D. 2008. A wide tuning range (1 GHz-to-15 GHz) fractional-N all-digital PLL in 45nm SOI. In Proc. 2008 IEEE Custom Integrated Circuits Conference (Sept. 21--24, 2008). CICC 2008. 431--434. DOI=10.1109/CICC.2008.4672113.Google ScholarCross Ref
- Rylyakov, A.; Tierno, J. A. 2010. Method and Apparatus for Low Latency Proportional Path in a Digitally Controlled System. U. S. Patent Application 2010/0017690 A1. January 21, 2010.Google Scholar
- Tierno, J. A.; Rylyakov, A. V.; Friedman, D. J. 2008. A Wide Power Supply Range, Wide Tuning Range, All Static CMOS All Digital PLL in 65 nm SOI. IEEE Journal of Solid-State Circuits. 43, 1 (Jan. 2008), 42--51. DOI=10.1109/JSSC.2007.910966.Google ScholarCross Ref
- Tiwari, A.; Torrellas, J. 2008. Facelift: Hiding and slowing down aging in multicores. In Proceedings of the 41st annual IEEE/ACM International Symposium on Microarchitecture (MICRO 41). IEEE Computer Society, Washington, DC, USA, 129--140. DOI=10.1109/MICRO.2008.4771785. Google ScholarDigital Library
Index Terms
- Active management of timing guardband to save energy in POWER7
Recommendations
Active Guardband Management in Power7+ to Save Energy and Maintain Reliability
Microprocessor voltage levels traditionally include substantial margin to ensure reliable operation despite variations in manufacturing, workload, and environmental parameters. This margin allows the microprocessor to function correctly during worst-...
High-speed and energy-efficient subthreshold level converter for wider voltage conversion
AbstractIn this paper, a high speed and energy-efficient dynamic threshold MOSFET (DTMOS) based hybrid level converter (DTHLC) is proposed with wide voltage conversion range from sub-threshold to nominal output voltage. The proposed level converter (LC) ...
An energy-efficient and ultra-low-voltage power oscillator in CMOS 65 nm
In a direct modulation scheme and particularly in portable and wireless sensor network applications, the oscillator limits the transmitter efficiency therefore, the oscillator efficiency is critical and high-efficiency and high-power oscillators are ...
Comments