All media appear to be going digital these days, including your photos, music, home movies, and even television. With the increasing ease of interacting with and storing this media, the storage space needed to hoard all this digital data is driving the production of even larger-capacity hard drives. When it's time to upgrade or add a hard drive in order to accommodate your growing media collection, you need to consider more than just gigabytes. It's not enough to just take into account storage capacity, you also must factor in drive performance. Not all drives are created equal, and you'll really feel the difference when you're waiting for those huge files to open. In order to deliver relevant performance evaluations of hard drives, CNET Labs puts drives through their paces with the real-world tasks of reading and writing files.

Test environment

Hard drives are tested using a desktop system with a 1.86GHz Intel Core 2 Duo E6300 Processor, 1GB of DDR2 SDRAM running at 667MHz, an integrated Intel Q965 Express graphics chip, a 250GB Western Digital Caviar SE hard drive, and Windows XP Professional SP2.

Internal SATA hard drives are connected to an available SATA II channel on the test bed motherboard. External SATA hard drives are connected to an available eSATA II port on a PCI-X controller card, which uses the Silicon Image Sil3124-2 controller chip. Internal IDE hard drives are the only installed IDE devices on the test bed's primary IDE channel; the drive being tested and the test bed's own SATA II hard drive are the only hard drives installed in the system.

External USB hard drives are connected to one of the USB 2.0 high-speed ports on the back of the test bed. Unless otherwise mentioned within the hard drive review, external FireWire hard drives are connected to the FireWire 400 or FireWire 800 ports on the back of the test bed. If an external drive supports both USB and FireWire connections, we will test the drive with whichever cable comes bundled with the drive, which is usually USB. If an external drive comes bundled with both USB and FireWire cables--or neither--we test using the USB connection.

All tested hard drives are formatted with a single NTFS partition at the maximum supported storage capacity of the drive. Tested drives will have only test files placed on them. The operating system and all application files reside on the test bed's permanent hard drive.

Transfer-speed tests

The read and write transfer-speed tests are done using a folder containing approximately 10GB of data that contains DOC, XLS, JPEG, GIF, HTML, TXT, MP3, AVI, and application installation files, ranging in size from 18KB to 698MB.

We use a custom utility designed by CNET Labs to simulate the drag-and-drop mode of file copying in Windows Explorer. The custom utility automatically times the file transfer tests, reporting how long it takes to complete the transfer in minutes and seconds as well as in megabytes per seconds.

Before testing begins, the test bed's permanent hard drive is defragmented using Windows' built-in defragmentation tool. The write test is conducted by timing how long it takes to copy the 10GB folder from the test bed's permanent hard drive to the drive being tested. The read test is conducted by timing how long it takes to copy the 10GB folder from the drive being tested to the test bed's permanent hard drive. All files copied to the drive being tested stay on the drive and are not deleted between test runs. This allows the drive to increasingly fill up with data as testing continues. The test bed's permanent drive, on the other hand, has all files copied to it deleted between test runs, in order to minimize the impact the test bed's drive will have on the performance of the drive being tested.

All tests are repeated a minimum of three times. Each reported score represents an average of three scores that are within 5 percent of each other; iterations that vary by more than 5 percent are thrown out. All scores are reported in minutes and seconds, so lower scores indicate faster performance.