Performance per watt

In computing, performance per watt is a measure of the energy efficiency of a particular computer architecture or computer hardware. Literally, it measures the rate of computation that can be delivered by a computer for every watt of power consumed.

The performance and power consumption metrics used depend on the definition; reasonable measures of performance are FLOPS, MIPS, or the score for any performance benchmark. Several measures of power usage may be employed, depending on the purposes of the metric; for example, a metric might only consider the electrical power delivered to a machine directly, while another might include all power necessary to run a computer, such as cooling and monitoring systems. Often the power is the average power used while running the benchmark, but sometimes other measures of power usage may be employed (e.g. peak power, idle power.)

For example, the early UNIVAC I computer performed approximately 0.015 operations per watt-second (performing 1,905 operations per second, while consuming 125 kW).

Most of the power a computer uses is converted into heat, so a system that takes fewer watts to do a job will require less cooling to maintain a given operating temperature. If installed where there is limited climate control, a lower power computer will operate at a lower temperature, which may make it more reliable. Reducing cooling demands make it easier to make a computer quiet. Reducing energy consumption can also make it less costly to run, and reduce the environmental impact from powering the computer (see green computing).

Computing energy consumption is sometimes also measured by reporting the energy required to run a particular benchmark, for instance EEMBC EnergyBench. Energy consumption figures for a standard workload may make it easier to judge the effect of an improvement in energy efficiency, just as the use of L/100km is easier than reciprocal measures (such as miles per gallon) when judging impact of automotive fuel economy.

Performance (in operations/second) per watt can also be written as operations/watt-second, or operations/joule, since 1 watt = 1 joule/second.

FLOPS per watt

FLOPS (Floating Point Operations Per Second) per watt is a common measure. Like the FLOPS it is based on, the metric is usually applied to scientific computing and simulations involving many floating point calculations.

Examples

Microwulf, a low cost desktop Beowulf cluster of 4 dual core Athlon 64 x2 3800+ computers, runs at 58 MFLOPS/Watt.cite web | url = http://www.calvin.edu/~adams/research/microwulf/power/ | title = Microwulf: Power Efficiency | author = Joel Adams | work = Microwulf: A Personal, Portable Beowulf Cluster ]

As of 2007, the Green500 list rates Blue Gene/P as the most efficient supercomputer on the TOP500 in terms of FLOPS per Watt, running at 350 MFLOPS/Watt. [cite web | url = http://www.baselinemag.com/c/a/Projects-Management/Super-Fast-and-Super-Green/ | work = Baseline Magazine | title = Super Fast and Super Green ]

On June 9th, 2008 CNN reported that IBM's Roadrunner supercomputer achieves 376 MFLOPS/Watt. [cite web | url = http://www.cnn.com/2008/TECH/06/09/fastest.computer.ap/index.html | work = CNN | title = Government unveils world's fastest computer ]

As part of Intel's TeraScale research project, they produced an 80 core CPU that can achieve over 16 GFLOPS/Watt. [cite web | url = http://www.tgdaily.com/content/view/30929/135/ | work = TG Daily | title = Intel squeezes 1.8 TFlops out of one processor] [cite web | url = http://techresearch.intel.com/articles/Tera-Scale/1449.htm | work = Intel Technology and Research | title = Teraflops Research Chip]

Green500 List

The Green500 list ranks computers from the TOP500 list of supercomputers in terms of energy efficiency. Typically measured as LINPACK FLOPS per watt. [cite web | url = http://www.green500.org | title = The Green 500] [cite web | url = http://www.itnews.com.au/News/65619,green-500-list-ranks-supercomputers.aspx | work = iTnews Australia | title = Green 500 list ranks supercomputers]

GPU efficiency

Graphics processing units (GPU) have continued to increase in energy usage, while CPUs designers have recently focused on improving performance per watt. High performance GPUs may now be the largest power consumer in a system. Measures like 3DMark2006 score per watt can help identify more efficient GPUs. [cite web | url=http://www.codinghorror.com/blog/archives/000662.html | title=Video Card Power Consumption | author= Jeff Atwood | date = August 18, 2006] However that may not adequately incorporate efficiency in typical use, where much time is spent doing less demanding tasks. [cite web | url = http://www.xbitlabs.com/articles/video/display/power-noise.html | title = Video card power consumption | work = Xbit Labs]

Since GPUs may also be used for some general purpose computation, sometimes their performance is measured in terms also applied to CPUs, such as FLOPS per watt.

Challenges

While performance per watt is useful, absolute power requirements are also important. Claims of improved performance per watt may be used to mask increasing power demands. For instance, GPUs have gotten to such large power requirements that though newer generation architectures may provide better performance per watt, they still consume unreasonable amounts of power. [cite web | url = http://www.bit-tech.net/columns/2007/05/20/performance_per_what/1 | title = Performance per What? | author = Tim Smalley | work = Bit Tech | accessdate = 2008-04-21 ]

Benchmarks that measure power under heavy load may not adequately reflect typical efficiency. For instance, 3DMark stresses the 3D performance of a GPU, but many computers spend most of their time doing less intense display tasks (idle, 2D tasks, displaying video). So the 2D or idle efficiency of the graphics system may be at least as significant for overall energy efficiency. Likewise, systems that spend much of their time in standby or soft off are not adequately characterized by just efficiency under load. To help address this some benchmarks, like SPECpower, include measurements at a series of load levels. [cite web | url = http://blogs.zdnet.com/Ou/?p=927 | work = ZDNet | title = SPEC launches standardized energy efficiency benchmark]

The efficiency of some electrical components, such as voltage regulators, decreases with increasing temperature, so the power used may increase with temperature. Power supplies, motherboards, and some video cards are some of the subsystems affected by this. So their power draw may depend on temperature, and the temperature or temperature dependence should be noted when measuring. [cite web | url = http://www.silentpcreview.com/article821-page5.html | title = Asus EN9600GT Silent Edition Graphics Card | author = Mike Chin | page = 5 | work = Silent PC Review | accessdate = 2008-04-21] cite web | url = http://www.silentpcreview.com/article814-page1.html | title = 80 Plus expands podium for Bronze, Silver & Gold | author = MIke Chin | work = Silent PC Review | accessdate = 2008-04-21 ]

Performance per watt also typically does not include full life-cycle costs. Since computer manufacturing is energy intensive, and computers often have a relatively short lifespan, energy and materials involved in production, distribution, disposal and recycling often make up significant portions of their cost, energy use, and environmental impact. [cite web | url=http://www.ecopcreview.com/LCA_and_ECPR | title=Life Cycle Analysis and Eco PC Review | work = Eco PC Review | author=Mike Chin] [cite journal | journal = Environ. Sci. Technol.
url = http://pubs.acs.org/cgi-bin/abstract.cgi/esthag/2004/38/i22/abs/es035152j.html
title = Energy intensity of computer manufacturing: hybrid assessment combining process and economic input-output methods | author = Eric Williams | year=2004
doi = 10.1021/es035152j
volume = 38
pages = 6166]

Energy required for climate control of the computer's surroundings is often not counted in the wattage calculation, but can be significant.cite journal | journal = CT Watch Quarterly
title = The Importance of Being Low Power in High Performance Computing
author = Wu-chun Feng
url = http://www.ctwatch.org/quarterly/articles/2005/08/the-importance-of-being-low-power-in-high-performance-computing/
volume = 1 | issue = 5 | year = 2005]

Other energy efficiency measures

SWaP (space, wattage and performance). SUN Microsystems metric for data centers, incorporating energy and space.

SW

Where performance is measured by any appropriate benchmark, and space is size of the computer. [cite journal | title = Sun SW
url = http://www.webhostingjournal.net/archives/Sun-SWaP-Metric-for-server-efficiency-and-datacenter-systems/#content | date = 6 december 2005 | journal = Web Hosting Journal ]

ee also

Energy efficiency benchmarks:
*Average CPU Power Measures power used running several standard benchmarks.
*SPECpower Benchmark for web servers running Java. (Server Side Java Operations per Joule)
*EEMBC EnergyBench.

Other:
*GeForce 8 Series GPU list, has energy use and theoretical FLOPS.

References

* cite journal | title = Making a case for Efficient Supercomputing
author = Wu-Chun Feng
journal = ACM Queue
volume = 1 | number = 7 | date = October 2003
url = http://www.acmqueue.com/modules.php?name=Content&pa=showpage&pid=80&page=1