These SSDs put a M.2 PCIe NVMe SSD behind a Thunderbolt 3 controller and deliver performance equivalent to an internal PCIe 3.0 x4 M.2 NVMe drive. External SSDs (Thunderbolt 3 interface): Consumers requiring very bus-powered external storage with very high performance are served by Thunderbolt 3 SSDs that are surprisingly affordable despite their performance profile.Such external SSDs now make up the bulk of the mid-range and high-end bus-powered external direct-attached storage devices. These have now evolved into M.2 SATA SSDs behind a much smaller and more power-efficient SATA - USB 3.1 Gen 2 Type-C bridge. External SSDs (USB interface): Over the last few years, we have had vendors place a 2.5' SSD board behind a SATA - USB 3.0 bridge.Most low-end and mid-range thumb drives have write speeds of the order of 20 - 80 MBps, and read speeds of the order of 100 - 150 MBps. Capacities have been steadily increasing, though very few high-end models boast of peak performance in the 300 - 400 MBps range. Thumb drives: These usually come with USB 2.0 or USB 3.0 interfaces, and are meant to plug in directly into the host port without an extra cable.On the external front, these devices fall under three major categories: Therefore, we obviously don’t have numbers for that.The advent of 3D TLC flash has enabled a large number of budget-friendly bus-powered solid-state storage devices. The MSI MEG X570 Unify does not support M.2 NVMe RAID. In any case, the results are excellent and there were no problems encountered during testing. The read speeds are actually slower than some of the ones I’ve seen with PCIe 3.0 based Samsung drives like the 970 EVO NVMe. Still, the testing accomplishes the goal of ensuring that the motherboard works properly.Īs we can see, the PCIe gen 4.0 interface allows the test drive to achieve outstanding write speeds. Actual performance is determined almost entirely by the device itself. NVMe drive testing is a little bit different than other drive types as it has very little to do with the bus provided the proper amount of PCIe lanes are available. USB 2.0 and 3.x came in at 37MB/s and 89MB/s respectively. We see a result of 295MB/s in the RAID 0 test and 126MB/s in the standalone AHCI test.

These are obviously not really tests designed to measure specific speeds on a given board, but rather to ensure that everything works as intended. Lastly, in the 100MB write test, we once again see typical results. USB 3.x and 2.0 achieved results of 303MB/s and 38MB/s. We saw a result of 581MB/s for the RAID 0 test and 125MB/s for the AHCI configuration. In the 100MB test set, we see results that are once again within expected ranges. USB 3.x and 2.0 results were 87MB/s and 38MB/s respectively. In this test, the RAID 0 configuration achieves better performance than the standalone AHCI configuration with 297MB/s compared to 121MB/s.

All of the devices perform in their expected ranges with no abnormalities. Once again, we see typical results for this platform. USB 3.x achieved a result of 310MB/s while USB 2.0 came in at 38MB/s which as I said before was expected. RAID 0 doesn’t really help here which is why it performs worse with this data set. The AHCI configuration achieved a result of 609MB/s while in RAID mode, the same drives achieve a result of 444MB/s. The results here are typical of what we see for the devices in question as well as motherboards utilizing this type of CPU and PCH.

Diskmark x