A new study just released by the California Energy Commission defies the conventional assumption that computers spend approximately half of their time switched on and the other half either in “sleep mode” using very little electricity or turned off altogether. This also means the estimates of all computer energy consumption globally and in the United States may be off – by a substantial amount.
In fact, findings from the report by the University of California Irvine (UCI) Plug Load Research Center imply that computer energy use in the United States could be higher than previously estimated by half or more. The study finds that office desktop computers are turned on as much as 77 percent of the time each day; and they are inactive 61 percent of that time – suffering insomnia, if you will, because they can’t go to sleep.
Although the study only monitored office desktop computers, these represent a large share of America’s computer energy consumption as desktops use on average five times as much energy as laptops and the large majority of desktops are in offices rather than homes. In addition, many of the issues identified by the study, such as user confusion about power management features exist for home desktop and laptop computers too.
If all computers in the United States used the same amount of energy as the ENERGY STAR™ benchmark for energy efficiency, the UCI study results would increase estimates of computer consumption in the United States from approximately 30 up to 45 billion kilowatt-hours annually. The higher number amounts to more electricity than all the households in New York City consume annually, is equivalent to the annual output of 17 large power plants, and costs American consumers and businesses $4 billion on their annual electricity bills.
What does this mean for you or me?
First, it’s a good reminder to manually switch off our computers and monitors when we’re done using them, which can save as much as $40 per computer on our annual electricity bills.
Second, it makes California Energy Commission’s (CEC) and the U.S. Department of Energy’s efforts to set efficiency standards for computers all the more important. Establishing energy-savings standards could lead to roughly $1.5 billion in U.S. utility bill savings, which will create jobs in California and nationally once those savings are reinvested in the local economy. Because the Golden State is home to one in every eight U.S. consumers, the minimum efficiency levels set in California are likely to have an impact on the energy use of computers sold across the nation, and they will pave the way for federal computer standards.
More on the UCI study
The study was based on a survey of 2,000 students and faculty at UCI using more than 3,300 computers, as well as software monitoring of the office desktop computers of a subset of 125 of the survey respondents. Interestingly, software monitoring found a very different picture from survey results, even for the same individuals: While 84 percent of participants had reported using automatic power management settings in the survey, only 20 percent of their desktops really had any automatic settings enabled.
Why such a difference between survey results and software monitoring?
The report points out survey participants may not understand or accurately remember power settings. Another possibility is that some computers may be correctly set to auto power-down, but applications are blocking the function, which is something manufacturers can’t control. Microsoft, Apple, and Google are best placed to help resolve this issue by making their operating systems less dependent on software applications using power management capabilities correctly.
Traditional power management is useful but can’t be the primary solution
For more than a decade, the ENERGY STAR program has required computers to be configured to go to sleep after 30 minutes or less of inactivity and the computer industry has voluntarily extended this configuration across all or most of their sales. The UCI study demonstrates that, sadly, this has not been effective. Some users change their settings, many corporate IT managers disable them, and even when enabled, some software applications prevent them from working properly. All of these barriers should be addressed. But realistically, the study suggests that power management cannot be relied upon to be the primary solution to computer energy efficiency.
“Keystroke Sleep” – A different kind of power management
Ironically, a different kind of power management seems to be winning the efficiency race. Some computer models, such as the Apple MacBook Pro, are designed to reduce processor power consumption at the millisecond level, including between keystrokes. This advanced real-time power management results in much lower average power use when the computer is in idle mode or performing low-intensity tasks such as browsing the internet.
The Apple MacBook Pro uses the same processor technology and chipset as a large number of models on the market (Intel Haswell platform). This shows that mainstream technology exists for computers to use very little power when performing no work. It’s time for this technology to be used on all computers, as it is on tablets and cell phones.
NRDC and the California investor-owned utilities have submitted a proposal for computer energy efficiency standards to the California Energy Commission, which only applies to idle, sleep, and off states, consistent with the ENERGY STAR version 6 energy efficiency requirements. Our proposal doesn’t set limits for active mode, such as surfing the internet, watching video or playing a game. Computers implementing keystroke sleep can dramatically reduce power in idle state, easily achieving our proposed idle limits.
What does this mean for CEC’s computer efficiency rulemaking?
UCI’s higher estimates of computer energy consumption mean that for each product unit, a given design improvement will save more energy. This makes the proposed standards more cost-effective, allowing the California Energy Commission to set more stringent efficiency requirements.
We urge the CEC to take into account the higher energy consumption estimates derived from the UCI study when it sets its efficiency standards for computers at technologically feasible and cost-effective levels.