USB 3.0 Speed Tests: 7-Way Host Controllers Roundup
We're going to go over the various host controllers quickly before we get to the test itself to give you an idea of some of the potential differences. This will be a fairly basic overview, as although there are architectural differences, much of that information isn't shared by the host controller makers. Sufficient to say there are at least a couple of different ways of doing things, as some manufacturers have split the PHY into USB 2.0 and USB 3.0 parts where each PHY handles both ports, whereas others have a single PHY per port for both USB 2.0 and USB 3.0.
The other main differentiator is xHCI (eXtensible Host Controller Interface) support, as some of the earlier designs only support xHCI 0.96; newer designs are following the xHCI 1.0 spec. As to exactly what has changed between version 0.96 and 1.0 is somewhat unclear as the previous version of the xHCI specification is no longer available. What we do know is that version 1.0 adds support for what is known as UX states which allows the various peripheral devices to go into a couple of low power sleep modes. Sadly very few devices have UX states enabled today, as it's not entirely backwards compatible with xHCI 0.96 and can cause disconnect problems. xHCI 1.0 is also meant to offer slightly improved performance, but as to how much, well, that we can't say, as it's impossible to test the difference due to it being an integral part of the chipset design.
All the manufacturers list which version of the xHCI standard their host controllers conforms to, so it's not hard to figure this one out. As for right now it makes very little difference in as much as we can tell if you get a 0.96 or a 1.0 xHCI compliant host controller, but ones power states of USB 3.0 devices start to be enabled en masse, there is a possibility of compatibility issues. Hopefully this is a problem that can be solved via a software workaround in some way.
There are also some ways to boost the performance of USB 3.0 storage devices by using software that "stretches" the standard to a degree. The easiest and least favorable way of doing this is to disable error checking, but this isn't done in any commercially available solution. For comparisons sake we did some tests with a motherboard from ASRock featuring their XFast USB software to see if this kind of software makes as much difference as it's claimed. The XFast USB software isn't going to cause you to lose any data, but it makes your USB 3.0 storage devices work outside of the USB 3.0 specification which you could in a way think of as overclocking the USB 3.0 bus.