Results: Tom's Hardware Storage Bench v1.0, Continued
Service Times
Beyond the average data rate reported on the previous page, there's even more information we can collect from Tom's Hardware's Storage Bench. For instance, mean (average) service times show what responsiveness is like on an average I/O during the trace.
It would be difficult to graph the 10+ million I/Os that make up our test, so looking at the average time to service an I/O makes more sense. For a more nuanced idea of what's transpiring during the trace, we plot mean service times for reads against writes. That way, drives with better latency show up closer to the origin; lower numbers are better.
Write latency is simply the total time it takes an input or output operation to be issued by the host operating system, travel to the storage subsystem, commit to the storage device, and have the drive acknowledge the operation. Read latency is similar. The operating system asks the storage device for data stored in a certain location, the SSD reads that information, and then it's sent to the host. Modern computers are fast and SSDs are zippy, but there's still a significant amount of latency involved in a storage transaction.
This is really funky, and I mean that in a good way. Silicon Motion's controller mated to Toshiba's Toggle-mode NAND become a potent force. The 128 GB drive achieves almost too-good-to-be-true read latency in our trace, and excellent write latency, too. A few other 128 GB-class SSDs are singled out in red, with the SM2246EN platform in blue.
Notice that Silicon Motion beats not just the 120 GB SanDisk Extreme II in our read latency measurement, but it also edges out Samsung's 840 Pro 128 GB by the slimmest of margins. Write latency is strong (in the same league as those two monsters), but far ahead of the M500 and 840 EVO, both leveraging 128 Gb flash.
The SM2246EN-based drive is basically tied with the 1 TB 840 EVO, one microsecond behind the Vector, and only six behind the 256 GB 840 Pro. That's not terrible company to keep. Remember, though, that this is a function of the controller and flash together. So, don't think that duct taping some triple-level cell NAND on the PCB would give you similar numbers.
And that takes us to another harsh reality. Regardless of whether you're using two- or three-bit per cell flash, migrating from 64 to 128 Gb density means you'll only need half as many dies at a given capacity. Unfortunately, this results in less parallelism for drives like Crucial's M500 and Samsung's 840 EVO. With those two SSDs, a higher-capacity model is definitely better.
