USB 3.0 Speed Tests: 7-Way Host Controllers Roundup
Believe us when we say that this is one of the most epic reviews we've ever put together. Not because it was the hardest roundup we've ever done, but it's by far the most time consuming one and it doesn't even have anything to do with the benchmarks we ran. We thought it'd be a good idea to do a comparative review of the various USB 3.0 host controllers out there, as by now we've finally reached a stage where there's some competition in the market with at least three major players and a couple of smaller ones. VR-Zone is also proud to have a world exclusive first review of the upcoming Renesas D720201 host controller which is launching later this year as part of this roundup.
To give a little bit of background information to the USB 3.0 standard we thought we'd do a quick rundown of the theoretical benefits over USB 2.0. As you most likely already know, USB 3.0 is promising over 10 times the bandwidth of USB 2.0 and although in reality USB 3.0 doesn't quite reach its theoretical bandwidth of 5Gbps. It is possible to reach speeds in excess of 400MB/s (3.2Gbps) and we've already seen this demoed, but and this is a big but, only if you use very specific hardware today. Currently we only know of one solution that can provide these kinds of speeds and it's a SATA to USB 3.0 RAID controller by Fujitsu, but to our knowledge it's only sold in one product made by a Japanese company.
On that premise, let's take a closer look at the contestants. The USB 3.0 host controllers we put through the paces are the AMD A75 chipset, the ASMedia ASM1042, the Etron EJ168A, the Fresco Logic FL1009, the NEC/Renesas µD720200, the Renesas µD720201 and the VLI VL800. Sadly we were unable to source a Texas Instruments host controller as the company requested to have an FAE present during the testing and this wasn't possible to arrange from their side.
Most of these host controllers can be found on a wide range of motherboards with the exception of the Fresco Logic FL1009 and the upcoming Renesas µD720201 which won't go into mass production until October. The integrated solution in AMD's A75 chipset was co-developed with Renesas and our understanding is that it's something of a hybrid between the older µD720200 and the upcoming µD720201.
The baseline for this test would have to be the NEC/Renesas µD720200 as it was the first USB 3.0 host controller on the market and it's become something of the de facto standard. As such we expected the other controllers to outperform it simply on the virtue of it having been developed so much earlier. However, we were surprised to see the old NEC controller keep up with the competition in most of the test we threw at it, although it's clear that there are faster options out there. One thing that helps the µD720200 stay ahead of the pack in some tests is native UASP (USB Attached SCSI Protocol) support in its drivers, which means that there's a performance boost to be had in specific tests thanks to the way the UASP drivers work compared to the standard BOT (Bulk Only Transfer) driver most of the controllers rely upon.