Measured Energy Savings and Performance of Power-Managed Personal Computers and Monitors

B. Nordman, M.A. Piette, K. Kinney

Power-management systems for the vast number of personal computers now in operation could help consumers save as much as $1 billion a year in energy costs. Personal computers (PCs) are estimated to use a total of 16 billion kWh/year of commercial sector electricity. The U.S. Environmental Protection Agency's Energy Star program specifies maximum power demand (30 W) for computers and monitors when in a "sleep"or inactive mode. To make better estimates of the electricity savings we can expect from power management, we gathered energy-use data from power-managed (Energy Star-compliant) PCs and monitors, audited the power-management features in a sample office area with 34 PCs, and reviewed the underlying technology that operates power management. In the audit, we examined whether the power-management features were "enabled," i.e., turned on.

We collected electricity-use measurements of eleven pieces of power-managed office equipment: three PCs and three monitors in LBNL offices, and five combined PC-monitor systems from several locations around the country. We measured the time the machines spent in primary system operating modes (off, low-, and full-power) and combined these with power measurements to derive hours of use per mode. Energy-savings estimates were derived using three schedules: "as-operated," "standardized" to a typical operating pattern, and the "maximum" achievable savings.

The Figure shows the as-operated and standardized savings for the eleven devices in annual electricity savings and dollar savings at the U.S. average commercial electricity price of 8¢/kWh. As-operated, energy savings for the eleven PCs and monitors ranged from 0 to 75 kWh/year; under the standardized operating schedule (device turned on 20% of nights and weekends), the savings are about 200 kWh/year. Most of the savings are from monitor power management. Measured separately, the monitors are physically larger than those in the combined systems, which explains why the savings for the two categories are similar. As is apparent from the Figure, monitor power management saved considerably more energy in these systems due to higher operating and lower "sleep" power levels.

Figure. Energy and dollar savings from PC power management, both as measured and under a standard operating pattern.

An audit of several dozen compliant devices found only 11% of PCs fully enabled and 39% of monitors properly configured. Further examination of the enabled devices showed that many were not entering low-power modes or were only powering down a few system components and thus achieving only small power savings. In some cases, we determined that the computer network hardware and activity was keeping the PCs "awake."

Changes in technology, rapid turnover of different computer models, and the interactions among hardware and software components make any characterization of the potential and actual state of power management in PCs complicated. Confusion among users and support personnel often leads to assumptions that power management is operating when it is not. Similarly, for some users, one bad experience with power management leads to permanent disablement of the feature.

The most critical outstanding research issue is to evaluate the percentage of compliant machines that are enabled and working. As of October 1995, Energy Star PCs and monitors were required to be shipped already enabled, which should increase the percentage of PCs and monitors using power management. Ensuring that these devices routinely enter low-power modes is also important.

The highest priority for increasing energy savings from power management is to target monitors since they are generally easier to configure than are PCs, are less likely to interfere with system operation, and have greater savings for each device. The difficulty of knowing how to configure most PC systems properly is the largest current barrier to achieving the savings potential from power management. Better software controls should greatly alleviate this problem.

Reference

Piette, MA, Cramer M, Eto J, Koomey J. Office Technology Energy Use and Savings Potential in New York. Lawrence Berkeley National Laboratory Report No. LBL-36752, 1995.

| Return to Contents | Return to Previous Paper | Next Paper |