How we tested: speeds and feeds

By changing the three critical components of our HP Compaq Presario V4000 test unit, we were able to produce configurations that differed significantly in performance and battery life. Our test bed used Intel's 900-series chipset.

Our test unit's CPU wasn't soldered on, so we were able to quickly swap it in and out. Though the V4000 can be configured with processors that range from an entry-level 1.4GHz Celeron to a top-of-the-line Pentium M, we settled on a midrange Pentium M 740 (1.73GHz) and an Intel Pentium M 770 (2.13GHz), two of the more popular configurations for this model.


The V4000 has two memory slots, giving you the option to use 256MB, 512MB, or 1GB RAM modules. We've found that 256MB of RAM is insufficient for running Windows XP without frequent crashes, so we used two Hynix PC2700S-25330 333MHz 512MB DDR SDRAM modules to configure the system with either 512MB or 1GB of RAM.

Hard drive options included three 4,200rpm models (60GB, 80GB, 100GB) and two 5,400rpm models (60GB or 80GB). We settled on two 80GB drives: a 4,200rpm Toshiba MK8025GAS and a 5,400rpm MK8026GAX.

Our V4000 test unit also contained these components:

• Windows XP Professional Edition with SP2 updates
• 15.4-inch wide-screen display at 1,280x800 resolution
• Intel 915 integrated graphics with up to 128MB of system memory
• Integrated 10/100Mbps wired LAN and 802.11b/g Wi-Fi
• 6-cell lithium-ion battery with 4,400 amp-hours of capacity
• 4X DVD burner
  
  

To gauge performance and battery life, we used BAPCo's MobileMark 2002 benchmark software, which runs nine real-world programs: Microsoft Word 2002, Microsoft Excel 2002, Microsoft PowerPoint 2002, Microsoft Outlook 2002, Netscape Communicator, McAfee VirusScan 5.13, WinZip 8, Macromedia Flash 5, and Adobe Photoshop 6.0.1. To simulate how people work, the benchmark creates sample documents, e-mail, and presentations and even takes occasional breaks. All the while, MobileMark 2002 measures how long it takes the battery to run down and how quickly the processor responds to user demands.

The results of our tests weren't particularly surprising: The faster the components, the better the performance--and the shorter the battery life. The highest-end configuration was 22 percent faster than the lowest-end configuration, and the most significant performance boost came from upping the RAM from 512MB to 1GB; the MobileMark 2002 score jumped from 171 to 195, though battery life dropped by 8 minutes.

Our least-effective configuration came from pairing a 2.13GHz Pentium M processor with 512MB of RAM and a slow 4,200rpm hard drive--clearly a mismatch of fast CPU and inadequate memory and hard drive. In fact, a configuration with a much slower 1.73GHz processor but 1GB of RAM and a 5,400rpm hard drive delivered superior performance and better battery life. The lesson: if you pay out for a fast processor, you'll need to pony up for other better components, too. And be prepared to trade speed for battery life; our highest-end system (2.13GHz, 1GB of RAM, 5,400rpm hard drive) got the highest MobileMark 2002 score (219) but had the shortest battery life (2 hours, 47 minutes).

More resources