In an interesting testament to differences between benchmarks,
IOMeter's hierarchy of drives is the reverse of that presented by
WinBench 99. That is, Intel's benchmark rates the
Cheetah 18XL as the fastest drive, the IBM trailing slightly behind, with the
Atlas 10k II lagging further still.
Under a Linear Workstation load (admittedly a scenario too simplistic to occur in real-world usage), the Ultrastar 36LZX pulls ahead of the Cheetah 18XL by a margin of nearly 4%. On the other hand, under Very Light and Light loads, IBM's offering trails that of Seagate's by 4%. The differences narrow under Moderate and Heavy loads... in Moderate cases, the Ultrastar trails the Cheetah by 1%. Heavy loads result in a dead heat.
The 36LZX does get to flex its muscles in IOMeter when compared to the Atlas 10k II. In such a face off, the Ultrastar leads the Atlas by margins in many cases approaching 8%.
Though our focus is primarily on Workstation performance, we also keep performance figures for File Server and Database scenarios. In such cases, the IBM drive lags behind the offerings from Seagate and Quantum. For full details, check out the StorageReview.com Database!
Conclusion
The UItrastar 36LZX continues the trend we've been witnessing towards quieter 10k RPM drives. Though its just a tad louder than the Cheetah 18XL (which, by the way, is an interesting reversal... IBM traditionally had the edge against the noisy Seagates), the 36LZX nonetheless brings relatively little seek noise to the table... certainly no worse than, say, the ATA Quantum Fireball Plus LM. Outside of a cooler, however, the drive operates quite warm to the touch in our midtower testbed. As should always be the case when dealing with 10k RPM drives, one should make sure the system case is well ventilated and/or active cooling is used.
Despite its better seek time, though, Big Blue's drive doesn't quite deliver the speed necessary to take the throne from Seagate. IBM's latest 10k RPM drive brings performance to the table just short of the Cheetah 18XL, currently the fastest 10K disk we've tested. As most would realize however when dealing with margins that don't exceed 4%, the differences are quite small and probably not noticeable in most cases. Since IBM and Seagate both share excellent reputations and manageable heat levels, we find ourselves reverting to an increasingly common recommendation: go with whatever you find to be the least expensive.