ABSTRACT
This paper reports and analyzes measured chip power and performance on five process technology generations executing 61 diverse benchmarks with a rigorous methodology. We measure representative Intel IA32 processors with technologies ranging from 130nm to 32nm while they execute sequential and parallel benchmarks written in native and managed languages. During this period, hardware and software changed substantially: (1) hardware vendors delivered chip multiprocessors instead of uniprocessors, and independently (2) software developers increasingly chose managed languages instead of native languages. This quantitative data reveals the extent of some known and previously unobserved hardware and software trends. Two themes emerge.
(I) Workload: The power, performance, and energy trends of native workloads do not approximate managed workloads. For example, (a) the SPEC CPU2006 native benchmarks on the i7 (45) and i5 (32) draw significantly less power than managed or scalable native benchmarks; and (b) managed runtimes exploit parallelism even when running single-threaded applications. The results recommend architects always include native and managed workloads when designing and evaluating energy efficient hardware.
(II) Architecture: Clock scaling, microarchitecture, simultaneous multithreading, and chip multiprocessors each elicit a huge variety of power, performance, and energy responses. This variety and the difficulty of obtaining power measurements recommends exposing on-chip power meters and when possible structure specific power meters for cores, caches, and other structures. Just as hardware event counters provide a quantitative grounding for performance innovations, power meters are necessary for optimizing energy.
Supplemental Material
Available for Download
The authors have provided experimental data and graphs associated with their article. The files are in .csv and xlsx formats.
- O. Azizi, A. Mahesri, B. C. Lee, S. J. Patel, and M. Horowitz. Energyperformance tradeoffs in processor architecture and circuit design: A marginal cost analysis. In ACM/IEEE International Symposium on Computer Architecture, pages 26--36, 2010. Google ScholarDigital Library
- C. Bienia, S. Kumar, J. P. Singh, and K. Li. The PARSEC benchmark suite: Characterization and architectural implications. Technical Report TR-811-08, Princeton University, January 2008.Google ScholarDigital Library
- W. L. Bircher and L. K. John. Analysis of dynamic power management on multi-core processors. In ACM International Conference on Supercomputing, pages 327--338, Island of Kos, Greece, 2008. Google ScholarDigital Library
- S. M. Blackburn, R. Garner, C. Hoffman, A. M. Khan, K. S. McKinley, R. Bentzur, A. Diwan, D. Feinberg, D. Frampton, S. Z. Guyer, M. Hirzel, A. Hosking, M. Jump, H. Lee, J. E. B. Moss, A. Phansalkar, D. Stefanović, T. VanDrunen, D. von Dincklage, and B. Wiedermann. The DaCapo benchmarks: Java benchmarking development and analysis. In ACM SIGPLAN Conference on Object-Oriented Programing, Systems, Languages, and Applications, pages 169--190, Oct. 2006. Google ScholarDigital Library
- S. M. Blackburn, K. S. McKinley, R. Garner, C. Hoffman, A. M. Khan, R. Bentzur, A. Diwan, D. Feinberg, D. Frampton, S. Z. Guyer, M. Hirzel, A. Hosking, M. Jump, H. Lee, J. E. B. Moss, A. Phansalkar, D. Stefanović, T. VanDrunen, D. von Dincklage, and B. Wiedermann. Wake up and smell the coffee: Evaluation methodologies for the 21st century. Communications of the ACM, 51(8):83--89, Aug. 2008. Google ScholarDigital Library
- M. Bohr. A 30 year retrospective on Dennard's MOSFET scaling paper. IEEE SSCS Newsletter, pages 11--13, Winter 2007.Google ScholarCross Ref
- K. Chakraborty. Over-provisioned Multicore Systems. PhD thesis, University of Wisconsin-Madison, 2008. Google ScholarDigital Library
- J. Charles, P. Jassi, N. S. Ananth, A. Sadat, and A. Fedorova. Evaluation of the Intel!R CoreTM i7 Turbo Boost feature. In IEEE International Symposium on Workload Characterization, pages 188--197, 2009. Google ScholarDigital Library
- J. S. Emer and D. W. Clark. A characterization of processor performance in the VAX-11/780. In ACM/IEEE International Symposium on Computer Architecture, pages 301--310, 1984. Google ScholarDigital Library
- J. S. Emer and D. W. Clark. Retrospective: A characterization of processor performance in the VAX-11/780. ACM 25 Years ISCA: Retrospectives and Reprints 1998, pages 274--283, 1998. Google ScholarDigital Library
- H. Esmaeilzadeh, S. M. Blackburn, X. Yang, and K. S. McKinley. Power and performance of native and Java benchmarks on 130nm to 32nm process technologies. In Sixth Annual Workshop on Modeling, Benchmarking, and Simulation, June 2010.Google Scholar
- H. Esmaeilzadeh, T. Cao, X. Yang, S. M. Blackburn, and K. S. McKinley. Source materials in ACM Digital Library for: Looking back on the language and hardware revolutions: Measured power, performance, and scaling. In International Conference on Architectural Support for Programming Languages and Operating Systems, Mar. 2011. Google ScholarDigital Library
- X. Fan, W.-D. Weber, and L. A. Barroso. Power provisioning for a warehouse-sized computer. In ACM/IEEE International Symposium on Computer Architecture, pages 13--23, San Diego, CA, 2007. Google ScholarDigital Library
- A. Georges, D. Buytaert, and L. Eeckhout. Statistically rigorous Java performance evaluation. In ACM Conference on Object-Oriented Programming Systems, Languages, and Applications, pages 57--76, 2007. Google ScholarDigital Library
- J. Ha, M. Gustafsson, S. M. Blackburn, and K. S. McKinley. Microarchitectural Characterization of Production JVMs and Java Workloads. In IBM CAS Workshop, Feb. 2008.Google Scholar
- M. Hempstead, G.-Y. Wei, and D. Brooks. Navigo: An early-stage model to study power-contrained architectures and specialization. In Workshop on Modeling, Benchmarking, and Simulations, June 2009.Google Scholar
- M. Horowitz, E. Alon, D. Patil, S. Naffziger, R. Kumar, and K. Bernstein. Scaling, power, and the future of CMOS. In Proceedings of International Electron Devices Meeting, pages 7--15, December 2005.Google Scholar
- Intel Corporation. Intel Hyper-Threading Technology, 2011. URL http://www.intel.com/technology/platform-technology/hyper-threading.Google Scholar
- Intel Corporation. Intel Turbo Boost Technology in Intel Core Microarchitecture (Nehalem) Based Processors. White Paper, Nov. 2008.Google Scholar
- C. Isci and M. Martonosi. Runtime power monitoring in high-end processors: Methodology and empirical data. In IEEE International Symposium on Microarchitecture, pages 93--104, December 2003. Google ScholarDigital Library
- ITRSWorking Group. International technology roadmap for semiconductors, 2011. URL http://www.itrs.net.Google Scholar
- E. Le Sueur and G. Heiser. Dynamic voltage and frequency scaling: The laws of diminishing returns. In Workshop on Power Aware Computing and Systems, Vancouver, Canada, Oct. 2010. Google ScholarDigital Library
- S. H. Li. Linux kernel bug 5471, 2011. URL https://bugzilla.kernel.org/show bug.cgi?id=5471.Google Scholar
- Y. Li, B. Lee, D. Brooks, Z. Hu, and K. Skadron. CMP design space exploration subject to physical contraints. In International Symposium on High Performance Computer Architecture, pages 17--28, Feb 2006.Google Scholar
- Nanoscale Integration and Modeling (NIMO) Group. Predictive technology model, 2011. URL http://ptm.asu.edu.Google Scholar
- V. Pallipadi and A. Starikovskiy. The ondemand governor: Past, present and future. In Proceedings of Linux Symposium, volume 2, pages 223--238, July 2006.Google Scholar
- Y. Patt. Future microprocessors: Multi-core, mega-nonsense, and what we must do differently moving forward. Distinguished Lecture at UIUC, (April 2010), 2011. URL http://www.parallel.illinois.edu/presentations/ 2010 04 30 Patt Slides.pdf.Google Scholar
- R. Sasanka, S. V. Adve, Y.-K. Chen, and E. Debes. The energy efficiency of CMP vs. SMT for multimedia workloads. In ACM International Conference on Supercomputing, pages 196--206, Malo, France, 2004. Google ScholarDigital Library
- R. Singhal. Inside Intel next generation Nehalem microarchitecture. Intel Developer Forum (IDF) presentation (August 2008), 2011. URL http://software.intel.com/file/18976.Google Scholar
- Standard Performance Evaluation Corp. SPEC Benchmarks, 2010. URL http://www.spec.org.Google Scholar
- Standard Performance Evaluation Corporation. SPEC CPU2006 benchmark descriptions. ACM SIGARCH Newsletter, Computer Architecture News, 34(4), September 2006. Google ScholarDigital Library
- The DaCapo Research Group. The DaCapo Benchmarks, beta-2006-08, 2006. URL http://www.-dacapo-bench.-org.Google Scholar
- TIOBE Software. TIOBE Programming Community Index for January 2011, 2011. URL http://www.tiobe.com.Google Scholar
- D. M. Tullsen, S. J. Eggers, and H. M. Levy. Simultaneous multithreading: Maximizing on-chip parallelism. In ACM/IEEE International Symposium on Computer Architecture, pages 392--403, Santa Margherita Ligure, Italy, June 1995. Google ScholarDigital Library
Index Terms
- Looking back on the language and hardware revolutions: measured power, performance, and scaling
Recommendations
Looking back on the language and hardware revolutions: measured power, performance, and scaling
ASPLOS '11This paper reports and analyzes measured chip power and performance on five process technology generations executing 61 diverse benchmarks with a rigorous methodology. We measure representative Intel IA32 processors with technologies ranging from 130nm ...
Looking back on the language and hardware revolutions: measured power, performance, and scaling
ASPLOS '11This paper reports and analyzes measured chip power and performance on five process technology generations executing 61 diverse benchmarks with a rigorous methodology. We measure representative Intel IA32 processors with technologies ranging from 130nm ...
Measuring and Benchmarking Power Consumption and Energy Efficiency
ICPE '18: Companion of the 2018 ACM/SPEC International Conference on Performance EngineeringEnergy efficiency is an important quality of computing systems. Researchers try to analyze, model, and predict the energy efficiency and power consumption of systems. Such research requires energy efficiency and power measurements, as well as ...
Comments