UltraDMA can mislead benchmarks
Rick Cook
The UltraDMA 33 hard disk interface offers significant performance improvements over older DMA versions in workstations and PCs. However the improvements aren't as great as some benchmarks suggest.
When UltraDMA 33 was introduced, the Internet newsgroups were flooded with reports of fantastic performance based on the results of common benchmarks.
On examination it turned out that the "performance" was an artifact of the drivers shipped with the new disks, not UltraDMA 33 itself. The most popular Windows benchmark for desktop systems, WinMark, turned the system memory caching off using the operating system command. The drivers shipped with some of the UltraDMA 33 disks overrode the operating system and turned system memory caching back on, artificially inflating the benchmark results by more than 250 percent in some cases. While UltraDMA 33 produced real performance increases compared to older DMA systems, they weren't nearly as impressive as the numbers yielded by the benchmarks.
The larger lesson is that benchmarks aren't the same thing as real-world performance and the best test of any piece of hardware and software's performance is on your system running your particular software. Beyond that, benchmarks of any sort need to be used with discretion and understanding of what's actually being measured and always to be taken with a grain of salt.
Rick Cook has been writing about mass storage since the days when the term meant an 80K floppy disk. The computers he learned on used ferrite cores and magnetic drums. For the last twenty years he has been a freelance writer specializing in storage and other computer issues.
Did you like this tip? If so (or if not) why not let us know? Follow this e-mail link to sound off, or visit our tips page to rate this tip and others.
