Monday, August 10th 2020

Device IDs of Intel's USB 4.0 Host Controller and Strategy to Phase Out Type-A Revealed

Device IDs of Intel's upcoming line of standalone USB 4.0 host controllers leaked to the web, courtesy Hardware Leaks (@_rogame). The controller possibly comes in three variants, bearing device IDs 0x9A1B, 0x9A1D, and 0x9A13. The alleged Intel confidential document screengrab speaks of USB 4.0 and USB 3.2 support (no mention of USB 2.0/1.1), and USB Power Delivery 3.0.

With USB 4.0, the USB-IF (USB Implementers Forum, or the special interest group behind USB), appears to want to standardize the USB type-C connector, eventually phasing out the type-A connector. To that effect, the document leaves out mention of USB 2.0/1.1 backwards compatibility. USB 4.0 debuts with an interface bandwidth of 40 Gbps, or 8 times that of USB 3.0, or over 80 times that of USB 2.0.
Sources: _rogame (Twitter), GenGoOn (image credit)
Add your own comment

38 Comments on Device IDs of Intel's USB 4.0 Host Controller and Strategy to Phase Out Type-A Revealed

#26
Assimilator
ncrsThis only happens if the USB 3.x controller can actually pass it down the chain. This is not always the case, for example this USB 3.1 A-Key to 3.0 converter cable requires a separate connection to a USB 2.0/1.1 controller:
That adapter is intended for use when your case has twin USB 3.0 ports on a single cable, but your motherboard only has a Type-C header. Since the USB 3.0 mobo header supports two ports, but the Type-C cable/header supports only one, the only option is to have two discrete cables to connect to the mobo. The manufacturer could've gone with two type-C headers, but no mobos have that many, so instead they went with a more usable split whereby you get one device running at 3.0 speeds and another running at 2.0 speeds. In the future, as more and more mobos go Type-C only, we will definitely see adapters of the former type.
ncrsWhile this probably will be the case it's not guaranteed by the USB 3.x specc.
Indeed, the most minimal/pure implementation of a USB 3.0-compliant controller will not support 2.0 or 1.1. But in practice there will always be a 2.0/1.1 controller behind the 3.0, or 4.0, or whatever.0 to fall back to.
Posted on Reply
#27
bonehead123
Die Type A, die, like, 'nuff already :D

This article is exactly why I continue to rag on any so-called "new" or "2020" cases that do not have C-type on the front I/O, even though almost every mobo made has had at least 1 on the back plane for quite some time now...

And pleeze don't give me that nonsense about how "C connectors cost so much more than A" ..... this is B.friggin.S....

if there were more of them being used/ordered by the mfgr's, the production of the connectors, cables, and peripherals would quickly get ramped up to meet the demand, and the costs would go down very quickly... just like virtually every other piece of tech component in the past 40 years or so :clap:

And >YES< I need moar of them, since I have multiple USB-C devices that I use everyday with my 'puter...
Posted on Reply
#28
Octopuss
I was very happy when I learned I could order a new type of connections panel for my Fractal case with all USB3 and USB C rather than mixture of 2+3.
Posted on Reply
#29
Frick
Fishfaced Nincompoop
OctopussYep, certain manufacturers are stuck in 1973 or something. There no reason whatsoever for COM being used for anything.
There absolutely is. It is super easy and dirt cheap to implement and it is easy to program for. Why make things complicated when there's no need?

And to clarify we are talking about RS-232. The COM port is a port.
OctopussUSB 2 should have died a horrible death few years ago.
I'm quite pissed off at motherboards still giving these useless connectors (well, technology, not the physical thing). Noone uses that crap for anything.
Do you mean the A ports? I would hate not having those connectors.

EDIT: Oh you mean USB 2.0. I still like to have one or two ports, I have found USB 3.0 to be a bit more sensitive. Some (old) stuff don't work with some USb 3.0 controllers.
Posted on Reply
#30
CrAsHnBuRnXp
OctopussYep, certain manufacturers are stuck in 1973 or something. There no reason whatsoever for COM being used for anything.
There is when the companies dont want to bother upgrading equipment.
Posted on Reply
#31
Wshlist
The thing is that a great many USB devices absolutely don't need more than USB2.0 speed. Now add the insane cost of USB4(+) cables, and their certification, and you are just being ridiculous.
I mean when something can be done for say $6 using a thin cable and instead you use a $25+ device + $100 cable solution you have to say you took a wrong turn somewhere.

I'm OK with USB-c (although I find it actually more fiddly to insert than the old large USB connectors I must say), but I'm not in favor of making everything USB4(+) compliant.
Posted on Reply
#32
Octopuss
I would just replace USB2 with USB3 on motherboards. There has to be evolution. USB4 looks just like PCIe 4 - just theoretical numbers that aren't good for anything real (ok, next gen NVMe SSDs in the case of PCIe 4, but that's not the usercase for most people...)
Posted on Reply
#33
Tardian
You can still get motherboards with DB25F parallel ports.
Posted on Reply
#34
king of swag187
R-T-BDoes apple even use USB-C? I don't think they do.
All computers since 2016 (2015 MB) and 2018/2020 iPad Pro lol
Posted on Reply
#35
R-T-B
king of swag187All computers since 2016 (2015 MB) and 2018/2020 iPad Pro lol
I meant in mobile, but should've specified. Doesn't matter they have them there too, lol.
Posted on Reply
#36
Ashtr1x
king of swag187All computers since 2016 (2015 MB) and 2018/2020 iPad Pro lol
Yep, And only those. I don't know how can people forget such moves by Apple where they just flat out remove all the ports to sell more of their accessories. MB only has USB C, iPad started to use USB C because of their marketing mentions how it works with cameras, extra device attachments, OTG stuff like that & compete with Surface and other Android tabs.

USB A phase out, I think will be only determined by market. USB A has more stronger port than USB C, the latter is too easy to disconnect. Not sure which port has higher wear and tear resistance, from my experience USB A has more resistance. Simply because of it's tight latching and how USB C cables are prone to kinking at the ends. USB C is expensive for these accessory makers I guess that's why no mouse uses it or a Mainstream RGB top KB like K95 use it either, and all of them use only USB 2.0 spec even DACs which are entry level use micro USB. Plus many of the phones still ship the same cables. It's robust and works. USB IF people are too stupid with their naming and standarization now it hit another peak.

Mobos still have PS/2 for those benchers and some of them allow Win7 installs too..
Posted on Reply
#37
ExcuseMeWtf
Vya DomusThey don't because the boards still support USB 2/1.1, see the problem ? Yes, you have to move on, if manufactures start releasing boards with just USB 3 you'll see devices supporting only USB 3 as well.
But why?

If it ain't broken, why fix it?

Do we really need "progress" for the sake of itself? Or is there some specific benefit those HIDs could bring using that extra bandwidth?
Posted on Reply
#38
Wshlist
ExcuseMeWtfBut why?

If it ain't broken, why fix it?

Do we really need "progress" for the sake of itself? Or is there some specific benefit those HIDs could bring using that extra bandwidth?
Does solve some confusion I suppose, at some point it becomes annoying to know if you have the right USB port if there are many kinds.

Problem is the chips having to dedicate a lot of bandwidth, and the solution would be 'flex USB', where the port would detect it's a USB2.0 device and then switch over to USB2 and free that excess bandwidth to other uses. Maybe put a small LED on it to indicate so, red for USB2 and blue for USB3 etcetera. (people need to be able to source issues.)
You could do that on the fly or on boot, depending on what's best for system designers.
To the manufacturers: Feel free to use that idea, I relinquish the rights for the good of the user.
Posted on Reply
Add your own comment
Oct 31st, 2024 20:38 EDT change timezone

New Forum Posts

Popular Reviews

Controversial News Posts