Traditionally, it's been StorageReview.com's policy to review the highest-capacity member of a given drive manufacturer's generation that retains a one-inch-high form factor. We've given these drives the moniker of "flagship"... we've contended that the performance delivered by a flagship unit is representative of the performance one can expect across the entire family. There have been some notable exceptions. IBM, for example, released its first 10k rpm drives solely in a 1.6" form factor and released the Ultrastar 18ZX well before the 9LZX, forcing us to take a look at the taller drive. We also made an exception for Hitachi's Pegasus, a 12,000rpm drive that was available only in a 1.6" version.
We've also decided to deviate from this policy when taking a look at Seagate's fourth-generation Cheetah offerings. Naturally, a look at the Cheetah 36LP was first. Shortly afterwards, we took a look at the Cheetah 18XL, a drive that not only featured a name different from the 36LP but also dissimilar performance (better, in fact... the 18XL rests in SR's Leaderboard listing). Next up is a look at the monstrous Ultra160/m SCSI Cheetah 73.
As its name implies, the 73 delivers an incredible 73.4 gigs of storage in a single drive, albeit in a 1.6"-high form factor. Though its scheduled to be bypassed amazingly quickly by the 75 gig ATA IBM Deskstar 75GXP, the Cheetah 73 is currently the largest single hard disk available. How does it differ from the Cheetah 36LP?
To achieve its enormous capacity, the 73 features 12 platters, each storing 6.1 gigs of data. The additional mass of the arms (doubled to 24 from the 36LP's 12) results in a slight increase in the drive's specified seek time- 5.6 milliseconds. Still no slouch, but decidedly mundane when there's drives on the horizon that dip below 5ms or even 4ms.
The 73's other specs remain identical to those of the 36LP and 18XL. As a Cheetah drive, the 73 features a 10,000rpm spindle speed. The unit comes equipped with a four meg buffer. It's backed by a five-year warranty.
In the past, we've attempted to gauge the effects of sequential transfer rates (nil in most situations) and buffer size (a mixed finding) on overall drive performance. The 73's increased seek time in a void of otherwise identical specs is probably as close as we can come to assessing this last quantifiable factor. Let's take a look at the results!
WB99/Win2k Low-Level Measurements