Sustained and scaling I/O rates
Our sustained IOMeter test hammers drives with 4KB random writes for 30 minutes straight. It uses a queue depth of 32, a setting which should result in higher speeds that saturate each drive's overprovisioned area more quickly. This lengthy—and heavy—workload isn't indicative of typical PC use, but it provides a sense of how the drives react when they're pushed to the brink.
We're reporting IOps rather than response times for these tests. Click the buttons below the graph to switch between SSDs.
To show the data in a slightly different light, we've graphed the peak random-write rate and the average, steady-state speed over the last minute of the test.
The Trion 150's peak speeds are on par with the Trion 100's. The newer drive's steady-state speeds are much improved, showing about a 75% increase over the the older model's. Even so, the Trion 150 is solidly in the slower half of our results. The 850 EVO's steady-state performance, for example, is half again as fast as the Trion 150's.
Our final IOMeter test examines performance scaling across a broad range of queue depths. We ramp all the way up to a queue depth of 128. Don't expect AHCI-based drives to scale past 32, though—that's the maximum depth of their native command queues.
For this test, we use a database access pattern comprising 66% reads and 33% writes, all of which are random. The test runs after 30 minutes of continuous random writes that put the drives in a simulated used state. Click the buttons below the graph to switch between the different drives. And note that the P3700 plot uses a much larger scale.
We never really expect budget SATA drives to exhibit much scaling. The database access pattern used for this test is certainly not the intended use case for a low-cost TLC drive like this one. Therefore, it's no shock that the Trion 150 doesn't seem to scale much more or less than the Trion 100 before it. IOps took a strange dive at QD4, but performance is roughly flat otherwise. The next graphs break down the results further and add some points of comparison. Click to toggle between read, write, and total IOps.
Something is definitely wonky here. Our write results demonstrate a weird bounce up and down as the test proceeds along from queue depth two to queue depths four and eight. We did a few extra runs of our scaling test, but got the same strange peaks and valleys each time. We've contacted OCZ to see if it can shed some light on what exactly is going on here, and we'll post an update if we get any answers.
That's it for synthetics. Up next is our suite of real-world performance tests. Let's see if the Trion 150 continues to distance itself from its predecessor.