Ever since the Cheetah's inception, Seagate has been careful to distinguish between families of its top-of-the-line drive. The original Cheetah 4LP, for example, was released simultaneously with the Cheetah 9, a drive that packed twice the capacity by doubling its platter count and thus shipping in a larger (1.6" high) package. Similarly, the second-generation Cheetah 9LP, consisting of 4.5 GB and 9.1 GB models, shipped along with the 1.6" Seagate Cheetah 18. The third-generation brought us the Cheetah 18LP, available in 9.1 and 18.2 gig capacities, with the Cheetah 36.
Each generation delivered us drives with a 1" form factor that performed similarly along with a 1.6" high drive designed with massive capacity in mind. The high capacity models featured a few drawbacks. The first, obviously, was the increased physical size. Such large form factors are more difficult to integrate... in terms of space, of course, but also cooling concerns. The second drawback was a slight performance hit: the increased platter count resulted in heavier actuator arms that couldn't access data quite as fast as its smaller, more nimble relatives.
In some ways, Seagate's fourth-generation Cheetahs mirror the strategy of the past. The flagship low-profile model, the Cheetah 36LP, is a drive featuring 36 gigs of capacity. The Cheetah 73, on the other hand, sacrifices small size for massive capacity. When it comes to the former, there's a unit featuring 36 gigs of capacity... and... that's it. Interestingly, there are no 18.2 or 9.1 GB Cheetah 36LPs. Such capacities have instead been given the fourth-generation moniker of "Cheetah 18XL." Why the difference?
Differences between the 18XL and the 36LP aren't discernable through the specs. Both drives feature 10,000rpm operation. Both units pack 6.1 gigs of data per platter. They both have a four megabyte buffer. Finally, each is rated with a seek time at 5.2 milliseconds.
Our contact at Seagate, however, is convinced that a measurable difference exists between the two lines. Apparently, in typical drive family situations, lower capacity drives that feature a correspondingly lower platter count don't necessarily feature lighter actuator arms. For instance, the 9 gigabyte Cheetah 18LP, despite featuring only three disks, is equipped with the same actuator assembly as the larger 18 gig model- one that features 12 arms. Thus, there really isn't any savings when it comes to smaller units. This is the principal reason that, contrary to popular belief, smaller capacity drives within the same family do not perform any better (in fact, they lag very slightly) than the flagship unit.
The Cheetah 18XL and 36LP, however, feature physically different actuator assemblies. The 36LP, featuring 12 platter-sides of data, is equipped with 12 arms. The 18XL, sporting only 6 platter-sides, comes with a correspondingly reduced assembly, featuring only 6 arms. Thus, despite the identical specified seek times between the two drives, the 18XL in theory should seek faster due to its lighter assembly. Does this pan out? Read on!
The Cheetah 18XL marks the first drive to be benchmarked solely in our new testbed. Our older system, finally, has moved on to not-necessarily-greater, not-necessarily-better things (its going to become my personal Linux experimentation box). It is our hope that the twenty-four ATA and SCSI drives that have been tested in both units provide sufficient overlap to interpolate the increasingly-irrelevant scores that a current drive may have posted on our older system. Now then, let's move on to some figures!
WB99/Win2k Low-Level Measurements