Followers of PC Perspective have likely seen a pair of stories previewing the upcoming performance and features of USB 3.1. First we got our hands on the MSI X99A Gaming 9 ACK motherboard and were able to run through our very first hands-on testing with USB 3.1 hardware. The motherboard had built-in USB 3.1 support and a device that was configured with a RAID-0 of Intel SSD 730 Series drives.
We followed that up with a look at the ASUS USB 3.1 implementation that included a PCIe add-on card and a dual-drive mSATA device also in RAID-0. This configuration was interesting because we can theoretically install this $40 product into any system with a free PCI Express slot.
Performance was astounding for incredibly early implementations, reaching as high as 835 MB/s!
In that last article I theorized that it would be some time before we got our hands on retail USB 3.1 hardware but it appears I wasn't giving the industry enough credit. ASUS passed us a list of incoming devices along with release schedules. There are 27 devices scheduled to be released before the end of April and ~35 by the middle of the year.
It's a daunting table to look at, so be prepared!
|
The product categories are mostly dominated by the likes of the a USB 3.1 to 2.5-in adapter; that would be useful but you aren't going to top out the performance of the USB 3.1 with a single 2.5-in SATA device. Iomaster has one listed as a "USB 3.1 to MSATA & M2 SSD enclosure" which could be more interesting - does it accept PCI Express M.2 SSDs?
Minerva Innovation has a couple of interesting options, all listed with pairs of mSATA or M.2 ports, two with Type-C connections. What we don't know based on this data is if it supports PCIe M.2 SSDs or SATA only and if it supports RAID-0.
A couple more list dual SATA ports which might indicate that we are going to see multiple hard drives / SSDs over a single USB 3.1 connection but without RAID support. That could be another way to utilize the bandwidth of USB 3.1 in a similar way to how we planned to use Thunderbolt daisy chaining.
We don't have pricing yet, but I don't think USB 3.1 accessories will be significantly more expensive than what USB 3.0 devices sell for. So, does this list of accessories make you more excited to upgrade your system for USB 3.1?
Ryan, Many thanks for your
Ryan, Many thanks for your recent coverage of USB 3.1; and, congratulations to the USB standards group for recognizing the advantages of the “jumbo frame” and higher clock rate that are already features of the PCI-Express 3.0 spec.
Because your website is a technical leader, I would like to see more discussion of changes that need to occur with the “SATA-IV” standard.
It’s clear now that the 6G clock rate and 8b/10b “legacy frame” are both obsolete and restraining progress with solid-state storage.
And, my reading confirms plans for PCI-Express 4.0 to adopt a 16G clock rate! Quoting pcisig.com:
“At 16GT/s bit rate, the interconnect performance bandwidth will be doubled over the PCIe 3.0 specification, while preserving compatibility with software and mechanical interfaces.”
Storage needs to keep pace with PCIe progress.
P.S. How about a future article about progress with Non-Volatile DRAM too?
KEEP UP THE GOOD WORK!
MRFS
What about PCI-SIG’s external
What about PCI-SIG’s external PCIe cable standard, are there any adopters outside of the industrial control market? If the USB folks want to really compete with TB, then the USB 3.1 controller chips better come in dual USB 3.1 variety for laptops, I’m not looking for lots of USB ports sharing one controller’s bandwidth. And a complete benchmarking and review of all the USB 3.1 controllers on the market, at least a good amount of the ones used in PC/Laptops is in order, and should be revisited/updated at least twice a year, to keep up with improvements. Laptop testing of USB 3.1 really needs to be considered also, not just on the testing rig PCI card based USB 3.1 controller, but get a broad cross section of laptops and do some benchmarking with their USB 3.1 offerings.
I see where you’re going with
I see where you’re going with that comment.
What of eSATA-IV oscillated at 8G and also used the 128b/130b “jumbo frame” now supported in PCIe 3.0?
Assuming no major changes to eSATA cabling, these two changes should produce a higher MAX HEADROOM of 8G/8.125b = ~985 MB/sec.
(130 bits / 16 bytes per PCIe jumbo frame = 8.125 bits per byte)
These 2 changes would also obviate the need for RAID-0 support in external SATA (“eSATA”) storage.
Add to this mix the existing “port multiplier” solutions for eSATA ports. Google “eSATA port multiplier” yields 34,400 hits today.
Lastly, looking forward to PCIe 4.0, such eSATA-IV+ features should allow for “auto detection” in order to detect and support a 16G transmission clock which is just around the corner.
MRFS
No the PCI-SIG standard is
No the PCI-SIG standard is faster than thunderbolt 2, and not as expensive, but the OEM market is so full of the undue influences of a few monopoly interests, and the backscratching that goes with it, that any external PCI open cabling standard gets the kibosh! Just wait until the USB Type-C standard plugs start being used in laptops, it is backwards compatible with USB 3.0/2.0 and I can see laptops getting the plug form factor standard long before the USB 3.1 controllers that will enable the 10Gbs bandwidth. laptops are the ones that benefit from the faster standards more, PCs with their available PCI expansion slots have more options, laptops have few, if any upgrade options, other than what the OEMs supply at purchase. This includes the lack of thorough testing that the gaming rig hardware gets, and most laptop “reviews” are little more than sponsored content to push the product, and make the “reviewer” a little scratch to supplement their script kiddy income.
> “the OEM market is so full
> “the OEM market is so full of the undue influences of a few monopoly interests, and the backscratching that goes with it, that any external PCI open cabling standard gets the kibosh!”
Very wise observation!
MRFS