Introduction and Background
We get hands on with the performance improvements of USB 3.1 due this year.
We first got a peek of USB 3.1 at CES 2015. MSI had a cool demo showing some throughput figures including read and write speeds as high as 690 MB/s, well over the ~450 MB/s we see on USB 3.0 options shipping today.
We were of course eager to play around with this for ourselves, and MSI was happy to oblige, sending along a box of goodies:
Stuff we will be testing today (Samsung T1 was not part of the MSI demo).
For those unaware, USB 3.1 (also known as Superspeed+), while only a 0.1 increment in numbering, incorporates a doubling of raw throughput and some dramatic improvements to the software overhead of the interface.
Don't be confused between the USB 3.1 standard and the new USB Type-C connector – they are unrelated and independent of each other.
Yes, you’re all going to have to buy *more* cables in the future.
Type-C connectors will enable more simple cable design and thinner connections going forward but USB 3.1 will exist in both Type-A/B and Type-C going forward. Our benchmarking today will utilize Type-A.
This is primarily a signaling change that a standard USB 3.0 port can handle. Don’t confuse that with the Type-C connector, which is essentially a different physical (invertible) packaging of the older connector we are all used to. This distinction is how we are able to test USB 3.1 speeds on the MSI X99A Gaming 9 ACK motherboard, which does not come equipped with a Type-C connector.
The various USB’s installed on the MSI X99A Gaming 9 ACK.
The little-reported-on nugget of info from the USB 3.1 specification relates to how they classify the raw vs. expected speeds. Taking USB 3.0 as an example, Superspeed can handle a raw 5Gbps data rate, but after subtracting out the overhead (packet framing, flow control, etc), you are left with ~450MB/s of real throughput. Superspeed+ upgrades the bit encoding type from 8b/10b (80% efficient) to 128b/132b (97% efficient) *in addition to* the doubling of raw data rate. This means that even after accounting for overhead, Superspeed+’s best case throughput should work out to ~1.1GB/s. That’s not a 2x speed improvement – it is actually 2.44x of USB 3.0 speed. Superspeed+ alright!
The older 8b/10b encoding, used on SATA (up to 3.1) and PCIe (up to 2.0), takes a 20% efficiency cut in throughput but was required for simpler PHY logic to transfer serial data without a clock. Newer logic uses a scrambler that accomplishes the same task far more efficiently.
To see what this new spec can do, we have a specialized demo board which uses an ASMedia ASM1352R, which is a USB 3.1 to dual SATA 6Gb/s (RAID) controller. The RAID is necessary as a single SATA SSD is not enough to saturate a USB link that can theoretically exceed 1GB/sec. On the other end of the link is the ASM1142 host controller, which is installed on the MSI motherboard used for this testing. Speaking of testing, let's see what these new ASMedia controllers are capable of.
usb blah blah blah, it’s
usb blah blah blah, it’s fast blah blah blah. so hows that 3440×1440 monitor.
It looks pretty sweet
It looks pretty sweet actually. We'll have something up on that shortly!
Allyn, Do you know if one
Allyn, Do you know if one could configure an OS “software” RAID using those 2 x AsMedia USB 3.1 integrated ports on the rear panel of that motherboard? You could eliminate that extra printed circuit board, if that were possible. I realize that one cannot boot from such a software RAID, but other OS functions could benefit e.g. pagefile.sys and browser caches. We have used a ramdisk with Firefox for several years, and the performance difference has been quite dramatic and consistent.
MRFS
You’re still going to see
You're still going to see much lower latencies with a good SATA or even PCIe SSD for those purposes. Too many layers between the host and the device with USB 3.1 to make it useful for those purposes.
Many thanks, Allyn:
YOU BE
Many thanks, Allyn:
YOU BE THE BEST!
p.s. Care to speculate about “SATA-IV”?
Seems logical to support the 128b/130b “jumbo frame”
in the PCIe 3.0 spec + a higher clock rate:
I like “interleaved” pre-sets: 6G, 8G, 12G, 16G etc.
(e.g. via auto-detection, jumper blocks, or Option ROMs).
Keep up the good work, Allyn.
/s/ Paul
See also:
Adata puts Type-C
See also:
Adata puts Type-C connector on tiny USB stick, speedy RAID drive
http://techreport.com/news/27627/adata-puts-type-c-connector-on-tiny-usb-stick-speedy-raid-drive
Those poor graphic
Those poor graphic designers…
Slaving away all day staring at monitors and ruining their eyesight to the point where they can’t notice that the they’ve made the 10Gbps bar more than 3 times longer than the 5Gbps bar.
If it stood for usable
If it stood for usable transfer rate, then 2.44x would be accurate, but it does look like it is about 3x in the graph.
Your calculation is very
Your calculation is very close. Here’s what I get:
USB 3.1 uses a 128b/132b jumbo frame with 16 bytes @ 8 bits each
i.e. 132 bits / 16 bytes = 8.25 bits per byte in each frame
10 Gbps / 8.25 bits per byte = 1212.12 Megabytes per second
USB 2.0 uses an 8b/10b legacy frame with 1 byte
i.e. 5 Gbps / 10 bits per byte = 500 Megabytes per second
1212.12 / 500 = 2.42X
HAHA – So true!
HAHA – So true!
Why USB 3.1 and not just USB
Why USB 3.1 and not just USB 4…? Just like why displayport 1.3a and crap why not just go to the next number…?