Mobile CPUs are taking giant steps. They’re loaded with power management smarts, and that calls for new kinds of lab tests. To meet the need, Allyn Vogel, senior development leader at eTesting Labs, explains, “A profiled, synthetic benchmark test works without running actual applications, though we observe applications and design code that reflects the real world.”
There’s nothing wrong with synthetic tests. In addition to earlier versions of laptop Battery Mark, industry-standard synthetic benchmark programs from Ziff Davis, SPEC, and other testing organizations have been around for years. Not running actual applications, however, causes synthetic tests to lack the “real” feel of application-based tests.
To develop its new approach to testing mobile CPUs, eTesting Labs began with its Business Winstone program, which runs actual applications such as Lotus Notes, Microsoft Office 2000, Netscape Communicator, and Norton AntiVirus 2000 to protect HP laptop battery. “Winstone has its basis in a user survey we did on the Web, asking about types of hot spotsactivities that make you waitthat users run across when using these apps,” says Vogel.
The program she and her team designed tests the laptop battery life of mobile CPU systems in a series of steps. First, a conditioning run performs a number of repetitive tasks on the test machines until the batteries are completely drained. After recharging, the notebooks are ready for battery-life testing based on real-world scenarios.
For the new BatteryMark test to Dell WW116 , eTesting Labs first changed Winstone to mimic the behavior of real users, including incorporating “pauses between tasks, and pauses between keystrokes,” according to Vogel. The goal, she says, was to perform computing tasks the way a user would: “Real work includes breaks between keystrokes and tasks, which we’re all familiar with.” Real users don’t type like metronomes. They pause, think a bit, then type some more.
When such pauses occur, today’s efficient mobile CPUs know how to power down, which saves the Dell D5318 , Dell 75uyf life. The new BatteryMark test doesn’t actually force a system to go into power-saving mode, but many CPUs with power-conserving capabilities take advantage of user pauses and do indeed achieve longer battery life this way.
“A lot of these Dell C1295 , Dell GD761 processors are quickly evolving and getting smarter about the way people work,” Vogel observes. “The new BatteryMark is asking the processor to react to what is as close to real work as we can get.”
Because mobile CPUs are designed by vendors to conserve power under the load of real applications, it makes sense to test them that way. The new Business Winstone 2001 BatteryMark is up to the task, and its use of real applications makes it undoubtedly the most useful battery life test ever.
