Vista's ReadyBoost flash drives lack significant boost

PC World - Windows Vista's Windows ReadyBoost sounds too good to be true, and based on our extensive lab tests, it is. The technology promises to let you speed up Windows by plugging an inexpensive USB flash drive into your PC. But we found that while ReadyBoost may speed up Vista a tiny bit, it can also slow it down in some instances.

The premise is this: Although writing data to and reading it from a flash drive is, in most cases, slower than writing and reading to a hard drive, if the data is scattered randomly in small chunks, then flash drives are faster. Vista's ReadyBoost is supposed to use that one speed advantage to create a faster, flash-drive-based cache of one of Windows' major bottlenecks -- the swap file on your hard drive that most Windows operations use. So ReadyBoost should theoretically speed up certain frequently performed Windows tasks such as loading programs.

The technology works with only the fastest flash drives -- those capable of 3.5MB/sec. throughput for 4KB random reads, and 2.5MB/sec. speeds for 512KB random writes. For this article, the PC World Test Center and I looked at three ReadyBoost-capable drives: Kingston's 1GB DataTraveler ReadyFlash, Lexar Media's 4GB JumpDrive Lightning and Ritek's Ridata 1GB Twister EZ Drive.

First, we used our WorldBench 6 Beta 2 benchmarking suite to test whether any of these devices sped up general Windows use when plugged into a desktop PC (an HP Compaq dc5750) and a notebook PC (an HP Pavilion tx1000). They didn't. At best, the Ridata had no impact on the notebook's WorldBench score. At worst, the Lexar slowed the desktop's WorldBench score from 42 to 39. Keep in mind that these scores represent only the subset of WorldBench 6 Beta 2 tests that we used for this story; we disabled the portions of WorldBench that interfered with ReadyBoost; as a result, several tests did not run. The HP Compaq dc5750 earned a full WorldBench 6 Beta 2 score of 62, while the HP Pavilion tx1000 earned a 64.

Painting a Target

We supplemented WorldBench by creating a test that was supposed to evaluate one of ReadyBoost's claims -- that it cuts the time required to load a program you've loaded many times before. Building a test around a product's purported strength is a bit like shooting an arrow into a wall and then painting a target around the arrow, but it was the best way to test this claim.

We learned that ReadyBoost does shorten the time it takes to load frequently used programs -- but not by much. The Lexar drive cut application load times by an average of 6% on our notebook and desktop PCs. Overall, we clocked launch-speed improvements of 4% to 6%. Without a stopwatch, you likely wouldn't notice the increase.