To measure the most basic performance parameters, we first start with a secure erased drive. We write over the entire LBA space with sequential writes twice, then write the capacity of the drive twice with 4K random writes. Once prepared, we run the following tests for one minute at each queue depth.
Throughput testing is the most basic kind of performance evaluation. After the requisite preconditioning on a freshly secure erased drive, we see that SMART wasn’t wrong about their 500MB/s read and write claims. Reads scale progressively from QD1 to QD64, from 180MB/s to 514MB/s. Writes are another story, hitting over 500MB/s at QD4 and peaking at 525MB/s. We’re constrained by the limitations of the 6gbps interface, but sequential performance is not the bread and butter of the Optimus.
4K read scaling looks good, peaking at 95K IOPS at a queue depth of 64. Average latency — the latency of each IO averaged throughout the test — remains minimal until reads stop scaling past 64 outstanding commands. At QD64, average latency is still only .687ms.
4K write scaling looks similar to the 128K scaling. That is, it peaks at relatively low queue depths. Latency is superb across the board, starting with .053ms at QD 1 (a scant 53 microseconds) and staying low until the writes stop scaling with QD when more than 32 commands are in the hopper. At a queue depth of 32, the Optimus pulls down 44K IOPS at an average latency of .741ms. Later, we’ll take a more in depth look at some write characteristics using a QD of 32.