The following post is a collection of excerpts from a draft of my Ph.D. depth paper, less citations, which exist in the original document.
Few computer systems spend all of their time at full utilization; even in always-on situations such as servers, a properly provisioned system will be spend almost all of its time at less than 50% utilization. Since most personal computers remain idle for extended periods of time ((In the context of this document, idle systems are ones that are on but performing no useful work.)), consumers should be considering a computer’s idle power draw when purchasing a computer. However, performance per watt is often the number that is compared.
[Studies suggest] that labeling products with energy consumption labels as is done with cars could have a positive impact on purchasing decisions. Currently, labels such as the different versions of the Energy Star certifications and proposed inclusion of performance per watt ratings are at best uninformative and worst misleading because the energy use of devices is highly dependent on usage patterns. Energy Star labels for computer equipment indicate that a particular component meets certain standards of efficiency when in standby mode (sometimes relative to their peak operating power) but, like performance per watt, give no indication of how much energy the component uses when idle or in use. Further compounding the problem is that CPUs are labelled with their thermal design power (TDP) in watts; not only do these numbers not give any indication about power draw when idle, the two major desktop CPU manufacturers, Intel and AMD, do not even use the same definition! The end result is that consumers generally do not know how much energy a device is consuming most of the time.
One step in the right direction is employed by camera manufacturers for measuring battery life. Standards put out by the Camera & Imaging Products Association (CIPA) suggest a way of estimating the battery life of a digital camera. However, because usage patterns vary to a large degree, a camera with a lower CIPA rating may out-last a camera with a higher one for one user and not another. Automobile fuel efficiency ratings take a different approach: they post fuel efficiency for highway driving and city driving. This is the computer equivalent of posting energy consumption for web browsing and playing games.
However, even this more detailed efficiency labeling does not always yield correct results. In the case of cars, automobiles consume fuel at different rates at different speeds and when accelerating and decelerating; city driving and highway driving fuel efficiency estimates are under the assumption that city driving involves more frequent stopping and that the car is traveling at a fast clip. On a congested freeway or a late night drive in the city, these assumption do not hold. Further, factors such as use of the air conditioning unit or increased aerodynamic drag from opening windows are not considered. Likewise, browsing animation-intensive websites or playing Minesweeper would run amok of the “browsing” and “gaming” labels.
One solution to this problem is to provide more or different information to the consumer (e.g., cruising fuel efficiency and acceleration/decelleration fuel efficiency or light-use efficiency and heavy use, though these terms themselves are problematic). While too much information may overwhelm and confuse consumers, insufficient information may be misleading. Being able to find a set of variables that is predictive and accessible could help reduce waste.