Results: 4 KB Random Performance
Again, we turn to Iometer as our synthetic metric of choice for testing 4 KB random performance. Technically, "random" translates to a consecutive access that occurs more than one sector away. On a mechanical hard disk, this can lead to significant latencies that hammer performance. Spinning media simply handles sequential accesses much better than random ones, since the heads don't have to be physically repositioned. With SSDs, the random/sequential access distinction is much less relevant. Data can be put wherever the controller wants it, so the idea that the operating system sees one piece of information next to another is mostly just an illusion.
4 KB Random Read
The Plextor, Seagate, and Corsair drives, all of which employ either 24 or 19 nm Toggle-mode NAND, regulate the field through a queue depth of 16. They're not matched until we push 32 commands concurrently.
OCZ's two drives fall into the middle of the pack, ahead of Intel's SSDs, but behind the three aforementioned models. It isn't until we hit a queue depth of 32 that the Vector and Vertex 450 pick up steam. At that point, the Vector puts 12,000 IOPS on the Vertex 450. With only a handful of outstanding commands, however, the delta is minimal. OCZ's Vertex 450 is rated for 85,000 4 KB random read IOPS, and we get slightly more.
4 KB Random Write
Roughly 10,000 IOPS separate the Vertex 450 from the Vector, but only at a queue depth of four. Everywhere else, the two drives perform virtually identically. Both SSDs are capable of writing faster at lower command counts than the other tested drives, though the entire field catches up at a queue depth of 32.
The Vector and Vertex 450 might have different names, but at least thus far, they aren't distinguishing themselves many other ways. After two 128 KB sequential writes across the entire LBA space, writing 4 KB blocks at a queue depth of 32 for multiple hours reveals the two SSDs as near-twins. The Vector is slightly faster and the Vertex 450 is slightly slower. Representing more familiar drives is Plextor's M5 Pro.
When we look at one-minute intervals, all three SSDs behave similarly. The Vector is still the quickest, but the Vertex 450 is consistently about 12% behind. Meanwhile, the Plextor manages fewer IOPS minute after minute. All three drives boast the same amount of flash, the same accessible space, and the same ratio of over-provisioning.
If we instead log at one-second intervals, not much changes except that the Plextor SSD exhibits a lower coefficient of variation. It's basically razor flat after shelving off what little space remains on a completely filled drive. The two OCZ models are quicker overall, though the spike higher and drop lower as they sort out their garbage collection situations.
None of these charts effectively show what's happening behind the scenes. In what is probably no coincidence at all, both Indilinx-based SSDs look a lot like the Marvell-powered Vertex 4 the Vertex 450 is supposed to replace. Clearly, we need some additional testing in order to help answer our questions.
