IOMeter — Sequential and random performance
IOMeter fuels much of our latest storage test suite, including our sequential and random I/O tests. These tests are run across the full capacity of the drive at two queue depths. The QD1 tests simulate a single thread, while the QD4 results emulate a more demanding desktop workload. For perspective, 87% of the requests in our old DriveBench 2.0 trace of real-world desktop activity have a queue depth of four or less. Clicking the buttons below the graphs switches between results charted at the different queue depths.
Our sequential tests use a relatively large 128-KB block size.
The XPG SX8200's sequential read speeds are respectable, right up there with Intel's NVMe drives. However, its sequential write speeds are lackluster for an NVMe SSD. We discussed these results with Adata, and the company explained that these speeds reflect the raw write speed of the 3D TLC NAND. In other words, the drive's pseudo-SLC caching mechanisms hadn't sufficiently recovered from writing the IOMeter test file to the drive and were effectively useless.
We've seen this sort of behavior with a few drives in the past. Some SSDs' caching implementations just don't play nice with our IOMeter test setup. We'll seek to address this when we update our storage bench and methods, but for now, results are results. The SX8200 will have plenty of opportunity to regain ground throughout the rest of our test suite.
The SX8200's random response times look great for both reads and writes. No caching shenanigans at play here.
The jury's still out on the SX8200. Random response times and sequential read speeds are just fine, but sequential write performance remains an open question. We'll revisit those again during our RoboBench tests. For now let's move on to sustained and scaling IOMeter tests.