SSD Benchmarks

Due to the inclusion of M.2 NVMe as a storage option, we also decided to run the full set of benchmarks using a Phison E18 NVMe SSD to see if there was any benefit of swapping the mechanical drives for an SSD. We used the same SSD for the SSD caching benchmarks, and it should be pretty straightforward to compare the numbers.

CrystalDiskMark

As you can see, the sequential read speeds are more or less identical across all different RAID levels, as well as when using an SSD, which simply means that the 2.5 Gbps Ethernet interface is the bottleneck. However, the random performance didn't improve when using an SSD, which suggests that there's some kind of bug on the OS or firmware level, as the performance shouldn't be this poor when using an SSD. The sequential write performance is roughly where it should be, so no issues there.

NAS Performance Tester

Once again, the 2.5 Gbps Ethernet interface is the limiting factor here, but the F4-424 edges out the Ugreen NASync DXP4800 Plus in when writing data to the NAS, despite having a less powerful SoC.

Helios LanTest

As with the previous tests, we're running into a bottleneck here as well, with the Ethernet interface limiting the data throughput.

Game Data

Oddly enough we saw somewhat lower performance using just an SSD, compared to using the same SSD for caching and even compared to the RAID 0 performance, using only hard disk drives. Even with a reduced PCIe bus width for the SSD, it ought to perform at least on par with the SSD caching test, but it falls behind slightly.

Single Large File

It seems like the SSD is better at handling a single large file, as in this specific test, the SSD comes out on top, if it's only by a smidgen. Again, a faster network interface would most likely have given the SSD a bigger advantage here.

Photos

Copying the photos to the SSD resulted in the worst performance out of all the tests we ran, which is really having us scratching our heads. As we run this test multiple times, it's not just a fluke either, but a repeatable result that makes no sense, especially as the SSD caching tests was almost 25 MB/s faster. The read speed was at the right end of the scale, but still being surpassed by two of the different SSD cache tests, if only by a slim margin.

It seems like TerraMaster is having similar oddities to Ugreen when it comes to adding an SSD to the NAS and this is something these companies need to look into.