USB Promoter Group Announces USB4® Version 2.0

[Doobeedoo]

Double the speed and already though. It's neat really. We need to start seeing more USB4 on new devices like mobos for sure, cases, external SSDs and USB sticks.

I'm just glad it's Type C only already. Really since USB4 is C only and base is USB4 and Version 2.0 doubles the speed I'm fine with it. Old ports need to die. But what is odd to me is why make USB4 have lower base speed than fastest older gen, rather unnecessary.

But yeah the more USB4 we see on new mobos the better. There should be more of them then old ones really.

[Eigenvektor]

1 hour ago, Doobeedoo said:

Old ports need to die. But what is odd to me is why make USB4 have lower base speed than fastest older gen, rather unnecessary.

I think you kind of answered your own question. They probably want to get rid of anything USB-non-4 and use USB4 for everything. You don't need USB4 80 Gbit to connect a mouse, but you probably also don't want to have to keep old USB protocols around forever. This way you can use one protocol for everything and adapt the speed to your use-case.

[Doobeedoo]

10 minutes ago, Eigenvektor said:

I think you kind of answered your own question. They probably want to get rid of anything USB-non-4 and use USB4 for everything. You don't need USB4 80 Gbit to connect a mouse, but you probably also don't want to have to keep old USB protocols around forever. This way you can use one protocol for everything and adapt the speed to your use-case.

Yeah I get that that you don't need it for peripherals though still odd to see the overlap in such way. While a mouse doesn't need 80Gbit or near it is understandable, though it's weird say base USB4 is only 10Gbit but older version could have twice of that, meaning the slowest USB4 can be half as fast in that case. Even protocols aside, they could've started with 20Gbit floor at least if not 40Gbit which would be faster than the old version.

[BigDamn]

3 hours ago, LAwLz said:

The reason why the same thing isn't happening with for example HP and AMD processors is because consumers understand that market better and would correctly aim their disapproval towards HP, not AMD.

This is naive, quite frankly. Not only are you giving the consumers a massive, unfounded, benefit of the doubt, but you are also assuming that vendors such as HP care at all about customers disapproving of their tactics. AMD provides consumer friendly product names that make sense, USB doesn't. If HP promoted their products using CPUs with technical names such as "8-Core Zen4" or "6-Core AM5", it comes off as unprofessional and unhelpful. The proper "Ryzen" nomenclature provides even the most uninformed consumers with adequate information regarding where easy CPU falls in the performance stack. The higher end products will sell themselves. Again, basic marketing.

3 hours ago, LAwLz said:

If AMD got the backlash for HP mislabelling processors, making them seem better than they are, then HP would do it for sure. Just like motherboard manufacturers and the likes misuse the USB names to confuse consumers and make their product seem better than they are.

This would likely be a contractual violation and/or trademark violation. AMD would NOT tolerate their own products being mismarketed and the negative PR with it. This is the correct response that any sensible company would have.

3 hours ago, LAwLz said:

The flaw isn't the USB-IF, the flaw is consumers blaming the USB-IF instead of blaming the manufacturers using incorrect terminology. We are starting to see the same trend with HDMI as well and it needs to stop. 

It isn't just manufacturers using incorrect terminology, it's everyone. Whens the last time you heard someone refer to USB 3.2 as "SuperSpeed++ USB 20gbps"? Probably never. And no, it isn't some giant conspiracy to make the USB-IF look bad, it's because that name is a huge pain in the ass to say. USB 3.2 is easy. Why would manufacturers refer to a naming scheme that no one uses? The USB-IF needs to reign this in, even though you compared this to being a "dictator".

[tikker]

1 hour ago, BigDamn said:

It isn't just manufacturers using incorrect terminology, it's everyone. Whens the last time you heard someone refer to USB 3.2 as "SuperSpeed++ USB 20gbps"? Probably never.

Because that's not the marketing name either. It's "SuperSpeed USB 20 Gbps".

1 hour ago, BigDamn said:

And no, it isn't some giant conspiracy to make the USB-IF look bad, it's because that name is a huge pain in the ass to say. USB 3.2 is easy. Why would manufacturers refer to a naming scheme that no one uses?

Why are people so against using what they're told the name of something is? Shorthand is fine, more when talking to a technical audience, but USB 3.2 does not mean what people think it means as USB-IF has stated on more than one occasion on record, in their presentations to the partners and manufacturers, and in their marketing guidelines. This has often been twisted into their naming scheme being confusing, but in reality is mostly just people misusing terminology that has been stated isn't what they should use. People complain that it's confusing that the 20 Gbps transfer mode is called USB 3.2 Gen 2x2, yet refuse to accept that there's an existing simple name for it that tells you exactly how fast it is: SuperSpeed USB 20 Gbps. I've stated multiple times that I agree they would benefit from enforcing nomenclature better, but there is most certainly at least equal blame on manufacturers simply not using the terms they are told to use.

[starsmine]

1 hour ago, tikker said:

Because that's not the marketing name either. It's "SuperSpeed USB 20 Gbps".

Why are people so against using what they're told the name of something is? Shorthand is fine, more when talking to a technical audience, but USB 3.2 does not mean what people think it means as USB-IF has stated on more than one occasion on record, in their presentations to the partners and manufacturers, and in their marketing guidelines. This has often been twisted into their naming scheme being confusing, but in reality is mostly just people misusing terminology that has been stated isn't what they should use. People complain that it's confusing that the 20 Gbps transfer mode is called USB 3.2 Gen 2x2, yet refuse to accept that there's an existing simple name for it that tells you exactly how fast it is: SuperSpeed USB 20 Gbps. I've stated multiple times that I agree they would benefit from enforcing nomenclature better, but there is most certainly at least equal blame on manufacturers simply not using the terms they are told to use.

Why doesnt usb 3.2 mean that is the point you are intentionally missing. 
Superspeed USB 20Gbps is an awful name for marketing. 

ALL usb IF had to do was make bigger number = better
3.0 into 3.1 into 3.2

the whole renaming 3.0 into 3.2 gen 1x1 or whatever it was was unnecessary and confusing.
When 3.1 Gen 2 is faster then USB 3.2 Gen 1, that just confuses people for no good reason.
Is USB 3.2 Gen 2 faster then USB 3.2? Nope. 3.2 is faster, why? consumers dont know.

USB IF just had to say "we defined the spec", and manufactures can then go, this port MEETS or EXCEEDS that spec (if they want to do their bullshit of adding half of a newer specs features)
No one knows the difference between Superspeed, High speed, and Full speed. Cause those all are fucking synonyms when used outside of this technical use case.
Also you are partially wrong on SuperSpeed USB 20Gbps (the correct way to type it sure), There IS "SuperSpeed+ USB 20 Gbps" there IS a + in there in the logo
 

So there you go, another inconsistency that will again throw people off for no good reason. 

[Jito463]

19 minutes ago, tikker said:

Why are people so against using what they're told the name of something is? Shorthand is fine, more when talking to a technical audience, but USB 3.2 does not mean what people think it means as USB-IF has stated on more than one occasion on record, in their presentations to the partners and manufacturers, and in their marketing guidelines.

It doesn't matter how much USB-IF states what the marketing terms should be, if they don't have any way to enforce it.  Going back to the previous comparison about HP and AMD, I'm quite sure that AMD has specific contractual obligations surrounding the marketing terms HP is and is not allowed to use.  This is what any responsible company would do.  If USB-IF doesn't want to get blamed for lousy marketing, then they need to take greater control over this and ensure all companies conform to the same terminology.  The current system only serves to engender confusion for the end users, and it's only going to get worse if they don't do something about it.

[BigDamn]

1 hour ago, tikker said:

Why are people so against using what they're told the name of something is?

Genericide. Why do people refer to a tissue as a Kleenex? (Common in most of the US, not sure about elsewhere). It's catchy, convenient, and common enough that it's become ingrained in peoples minds. Does Kleenex need to slap the word "tissue" on to the box to pay homage to the original idea of the tissue? Intelligent marketing teams will name their products in this way to ensure people immediately think of their product when applicable. "SuperSpeed USB 20gbps" does not adhere to this marketing strategy, even without the "++", which is part of the official logo.

 

1 hour ago, tikker said:

This has often been twisted into their naming scheme being confusing, but in reality is mostly just people misusing terminology that has been stated isn't what they should use. People complain that it's confusing that the 20 Gbps transfer mode is called USB 3.2 Gen 2x2, yet refuse to accept that there's an existing simple name for it that tells you exactly how fast it is: SuperSpeed USB 20 Gbps.

If it was so simple they wouldn't have these problems. People have had qualms with WiFi, HDMI, etc. at times, but no one has consistently remained under the gun from day one like USB has.

 

1 hour ago, tikker said:

I've stated multiple times that I agree they would benefit from enforcing nomenclature better, but there is most certainly at least equal blame on manufacturers simply not using the terms they are told to use.

I'm not going to blame manufacturers for using the easier name, but who we blame doesn't matter. We agree that enforcing the correct terminology is the proper way to handle this, and is an ACTUAL solution.

[wamred]

On 9/1/2022 at 3:54 PM, Eigenvektor said:

Summary

USB4 will receive a speed update (80 Gbps). Keeping to their tradition of braindead versioning, the USB Promoter Group has chosen to call it USB4 2.0.

 

Quotes

 

My thoughts

Do those guys (and gals) seriously not understand how versioning works? It's nice to see new tech, why use this completely weird naming though...

 

Sources

• https://www.businesswire.com/news/home/20220901005211/en/USB-Promoter-Group-Announces-USB4®-Version-2.0
• https://www.tomshardware.com/news/usb-4-version-2-announced-80gbps

No, they don't. Nothing about their naming system makes sense. You don't need to retroactively change the name of every technology when you make an updated or new one.

[Eigenvektor]

3 hours ago, Doobeedoo said:

Yeah I get that that you don't need it for peripherals though still odd to see the overlap in such way. While a mouse doesn't need 80Gbit or near it is understandable, though it's weird say base USB4 is only 10Gbit but older version could have twice of that, meaning the slowest USB4 can be half as fast in that case. Even protocols aside, they could've started with 20Gbit floor at least if not 40Gbit which would be faster than the old version.

The idea is to get rid of the overlap. Replace everything with USB4. You don't need 20 Gbit minimum to do that, 10 Gbit is more than enough for most peripherals.

 

Right now motherboards have controllers for USB 2.0, usually multiple versions of USB 3.x and possibly USB4 if they're new enough.

 

You could just use USB4 for everything instead. Replace USB 2.0 with USB4 10 Gbit, replace USB 3.x with USB4 20 Gbit, then include one or two USB4 40 and/or 80 Gbit ports for fast storage. As long as things are backwards compatible, there's no real reason to include USB 2/3 anymore.

[Jito463]

1 hour ago, Eigenvektor said:

As long as things are backwards compatible, there's no real reason to include USB 2/3 anymore.

The main reason for including USB 2.0 still is for backwards compatibility with older operating systems.  If you try to use an older OS, and it doesn't have the drivers for USB4, then you can't use your keyboard/mouse.  If there would be a way to mimic USB 2.0 compatibility, then it would be a moot point.  While the issue is fairly niche, it's still a legitimate reason.  It's part of the reason you still see single PS2 ports on newer motherboards.

[tikker]

  

3 hours ago, starsmine said:

Why doesnt usb 3.2 mean that is the point you are intentionally missing. 
Superspeed USB 20Gbps is an awful name for marketing. 

Why? The main complaints I see is that people loose track of what [speed] Gen X YxZ implies, which would never have been a confusion if manufacturers used the intended name that literally tells you the speed.

3 hours ago, starsmine said:

ALL usb IF had to do was make bigger number = better
3.0 into 3.1 into 3.2

All the consumer and manufacturers had to do was simply use the given speed nomenclature to indicate what the speed is, which automatically means bigger number == "better".

3 hours ago, starsmine said:

USB IF just had to, we defined the spec, and manufactures can then go, this port MEETS or EXCEEDS that spec (if they want to do their bullshit of adding half of a newer specs features)

They did. That spec is called "USB" and does not imply a set of features as you imply here. It tells you your options on how to build a USB device right now. If you advertise USB 3.2 that just means that you used version 3.2 of the specification to build the device just like if I build a house now I build it using the 2022 specification.

 

They have done as you say: they provide the spect and they have a set of terms that manufacturers can meet or exceed. Can it do 10 Gbps? You label it as 10 Gbps. Can it do 20 Gbps? You label it as 20 Gbps. It doesn't get simpler then that, but people shoot themselves in the foot on purpose and then complain it hurts.

3 hours ago, starsmine said:

No one knows the difference between Superspeed, High speed, and Full speed. Cause those all are fucking synonyms when used outside of this technical use case.

And is it a big issue to read two words more and see "SuperSpeed USB 20 Gbps" literally telling you that you have a 20 Gbps USB cable?

3 hours ago, starsmine said:

Also you are partially wrong on SuperSpeed USB 20Gbps (the correct way to type it sure), There IS "SuperSpeed+ USB 20 Gbps" there IS a + in there in the logo
 

So there you go, another inconsistency that will again throw people off for no good reason. 

SuperSpeed+ is not the marketing name and is not to be used in advertising:

Quote

https://www.usb.org/sites/default/files/usb_3_2_language_product_and_packaging_guidelines_final.pdf
USB 3.2 Gen 1
o Product capability: product signals at 5Gbps
o Marketing name: SuperSpeed USB
• USB 3.2 Gen 2
o Product capability: product signals at 10Gbps
o Marketing name: SuperSpeed USB 10Gbps
• USB 3.2 Gen 2x2
o Product capability: product signals at 20Gbps
o Marketing name: SuperSpeed USB 20Gbps

NOTE: SuperSpeed Plus, Enhanced SuperSpeed and SuperSpeed+ are defined in the USB
specifications however these terms are not intended to be used in product names,
messaging, packaging or any other consumer-facing content.

It is only in the logo, which I don't argue against being inconsistent. I never said the USB marketing was perfect. Only that in my opinion a substantial part of it does not inherently stem from the specfication.

3 hours ago, Jito463 said:

It doesn't matter how much USB-IF states what the marketing terms should be, if they don't have any way to enforce it.  Going back to the previous comparison about HP and AMD, I'm quite sure that AMD has specific contractual obligations surrounding the marketing terms HP is and is not allowed to use.  This is what any responsible company would do.  If USB-IF doesn't want to get blamed for lousy marketing, then they need to take greater control over this and ensure all companies conform to the same terminology.  The current system only serves to engender confusion for the end users, and it's only going to get worse if they don't do something about it.

As I said before, I agree that the USB-IF should be more active in helping achieve this. I don't agree with putting all the blame on USB-IF when they literally have guidelines that contain clearer terms then the ones people complain about. I know this is somewhat of an idealistic/utopian argument, but the companies wilffully ignore it and thus also deserve flack for it in my opinion.

 

2 hours ago, BigDamn said:

Genericide. Why do people refer to a tissue as a Kleenex? (Common in most of the US, not sure about elsewhere). It's catchy, convenient, and common enough that it's become ingrained in peoples minds. Does Kleenex need to slap the word "tissue" on to the box to pay homage to the original idea of the tissue? Intelligent marketing teams will name their products in this way to ensure people immediately think of their product when applicable. "SuperSpeed USB 20gbps" does not adhere to this marketing strategy, even without the "++", which is part of the official logo.

 

It was more of a rhetorical "why not use the easier terms as you are told" to the "why don't they do what I want". I think this is more similar to if Kleenex made, say, double and triple layered toilet paper internally labeleld as Kleenex TP A and Kleenex TP B, you and manufacturers using that instead of the official name and then complaining that the given "Kleenex Toilet paper Double-layered" is too long and cumbesome while also complaining it's unclear what A and B mean. If you got asked how fast your car goes you would answer it can do 148 km/h, not that it has a 7.3 liter V8.

 

You raise an important aspect though: "to ensure people immediately think of their product when applicable". I do understand why people gravitate towards HDMI 2.1 or USB 3.2. They want a single word with maybe one or two numbers to describe everything their product offers, but that is the thing: USB isn't a product. it's a standard. It describes a set of connectors and communication and power protocols. Saying "this is a Kleenex tissue" tells you it's a tissue made by Kleenex. It doesn't tell you what features of the International Tissue Standard Kleenex decided to implement. Similarly saying "this is a USB 3.2 device" tells you it's a device built against the version 3.2 of the USB standard, that's it. USB is the manual that Kleenex references on how they can make their tissues. Not what they advertise on the box.

 

I think it's fine to argue about the length of the term and SuperSpeed/HiSpeed. USB-IF seems to agree, as the marketing terms for USB4 seem to be just "USB4 20/40/80 Gbit/s", making this a nonissue for USB4 devices. I do not think that a 20 Gbit/s label conveys the speed less efficiently than a "3.2".

2 hours ago, BigDamn said:

If it was so simple they wouldn't have these problems. People have had qualms with WiFi, HDMI, etc. at times, but no one has consistently remained under the gun from day one like USB has.

Wi-Fi had the benefit that the underlying standard (IEEE 802.11) has already a complicated-looking name (historically, from what I read). They made a nice decision introducing the Wi-Fi 4, 5 and 6  branding, but then messed it up again by introducing 6E and 6 Version 2. HDMI also got flack and is the same misunderstanding as USB and does not approve pure-version number advertising either. HDMI cables should be labeled "Ultra High Speed HDMI Cable", for example, and for the rest you need to specify which feature you implement. Despite that, both standards got reported as "rebranding" multiple times. This shows that even tech outlets do not always fully understand what these versions actually refer to.

 

Just to be clear, I'm not saying the naming is perfect.

 

Why USB got more flack, I think, is that USB's changes focussed on technical aspects that bear names like "Gen 2x2", which are not familiar to us since they are USB-specific. HDMI got more every-day-life features such as "Variable Refresh Rate (VRR)" and "Auto Low Latency Mode (ALLM)" which were familiar terms. I don't think the HDMI specification is publicly available either, while the USB spec is, so the latter will reach a much broader audience for better of for worse.

 

1 hour ago, wamred said:

No, they don't. Nothing about their naming system makes sense. You don't need to retroactively change the name of every technology when you make an updated or new one.

Look at it this way: if you release a bunch of fixes and new features for your software package, don't you update the version number? That is what the USB versioning is. It versions the document telling you how to build a USB device.

[Eigenvektor]

20 hours ago, Jito463 said:

The main reason for including USB 2.0 still is for backwards compatibility with older operating systems.  If you try to use an older OS, and it doesn't have the drivers for USB4, then you can't use your keyboard/mouse.  If there would be a way to mimic USB 2.0 compatibility, then it would be a moot point.  While the issue is fairly niche, it's still a legitimate reason.  It's part of the reason you still see single PS2 ports on newer motherboards.

I get what you're saying, but we're probably talking a few more years into the future before there's an "everything is USB4" board.

 

An OS that is old enough at that point to no longer receive USB4 support likely doesn't really support SATA or NVMe either. I would question whether you really need a modern board for such a system. I'd be in favor of deprecating the stuff and if you really need an old OS, combine it with old-ish hardware. Alternatively, there will probably still be boards around that include at least one USB 2.0 port.

[infigo]

On 9/3/2022 at 6:04 PM, tikker said:

Oh yeah, I'm on board with the 20/40/80 Gbps labeling. I meant that I don't think they'll go back to "USB 4.1" or something like that, but will stick to USB4 [Version] X.Y for the spec.

The problem with that notion is that usb 4 20 gigabit may be hard to know which 20 gigabits. 

The usb3 gen2x2 version using 2 pairs of gen 2 or a single lane "gen3" from the new usb4 specification which only requires 1 pair and then allows for 4k display support with dp2.0 or hdmi on the other pair in the same cable. 

 

So for simple people, using usb something X speed gigabit might be easier, it also hides all features the devices support. 

 

But then again, people don't understand hdmi either so it might be a necessary evil. 

 

Allways an issue with specifications which can do a lot if things, it's either confusing (complicated) or too simplified...which causes different confusion when things are assumed to work when it is not that simple....case and point hdmi "versions" which simplified branding too much. 

[Eigenvektor]

1 minute ago, infigo said:

The problem with that notion is that usb 4 20 gigabit may be hard to know which 20 gigabits.

It's probably not going to matter to most people whether their USB4 20 Gbps port is actually USB4 Gen 2×2 or Gen 3×1, as long as it works as advertised.

 

If the slot is SuperSpeed USB 20Gbps (USB 3.2 Gen 2×2) instead, that should definitely still be distinguished from USB4, since it's not the same, even if it has the same speed. My personal favorite would be to label it as USB3 20 Gbps instead, to use the same naming conventions as USB4 while clearly marking it as an older version.

[infigo]

4 minutes ago, Eigenvektor said:

It's probably not going to matter to most people whether their USB4 20 Gbps port is actually USB4 Gen 2×2 or Gen 3×1, as long as it works as advertised.

 

If the slot is SuperSpeed USB 20Gbps (USB 3.2 Gen 2×2) instead, that should definitely still be distinguished from USB4, since it's not the same, even if it has the same speed. My personal favorite would be to label it as USB3 20 Gbps instead, to use the same naming conventions as USB4 while clearly marking it as an older version.

But current usb 4 usb 20 gigabits is still usb 4. 

Usb4 CAN use thunderbolt for more speed and driect PCIe connection...but don't have to, can still use usb as protcoll  which is identical to usb 3.2 gen2x2....but it is usb4. 

So we are already passed that notion. 

 

[Sauron]

Cool but I struggle to see a use case for it today other than splitting it in a hub. Barely anything can make good use of that much bandwidth even among internal components.

 

I don't care about the name, I just wish manufacturers would just state the bandwidth of the port and get it over with

[tikker]

3 hours ago, infigo said:

The problem with that notion is that usb 4 20 gigabit may be hard to know which 20 gigabits. 

The usb3 gen2x2 version using 2 pairs of gen 2 or a single lane "gen3" from the new usb4 specification which only requires 1 pair and then allows for 4k display support with dp2.0 or hdmi on the other pair in the same cable.

The entire point is that you as a consumer generally don't need to know whether it is electrically dual or single-lane. All you need to know is that it can do 20 Gbit/s and the device will internally figure out how to do that. USB is made to be backwards compatible, so a newer device should be able to accomodate an older device or transfer mode easily. The example you give is also one of the main points of USB4, so I think that should also be a non-issue, since that is part of the design. From the spec:

Quote

The difference with USB4 versus USB 3.2 is that USB4 is a connection-oriented, tunneling
architecture designed to combine multiple protocols onto a single physical interface, so that the
total speed and performance of the USB4 Fabric can be dynamically shared. USB4 allows for USB
data transfers to operate in parallel with other independent protocols specific to display,
load/store and host-to-host interfaces. Additionally, USB4 extends performance beyond the 20
Gbps (Gen 2 x 2) of USB 3.2 to 40 Gbps (Gen 3 x 2) over the same dual -lane, dual-simplex
architecture.

If, in your example, the cable/device can sustain 20 Gbit/s transfer in addition to other devices that means it can do more than 20 Gbit/s and thus  it should likely be a 40 Gbit/s cable, for example, in the first place. That would make it irrelevant again, to the consumer, how that 20 Gbit/s is achieved.

3 hours ago, infigo said:

So for simple people, using usb something X speed gigabit might be easier, it also hides all features the devices support. 

Do you have an example where you explicitely had to know what the supported transfer modes were? The feature is X Gbit/s transfer speeds, not a specific transfer mode. The spec states that any port supporting USB 3.2 connected to a device supporting USB 3.2 should support the USB 3.2 SuperSpeed and SuperSpeedPlus modes, which I think implies all the transfer modes should be supported anyway. That would make this a non-issue.

3 hours ago, infigo said:

But then again, people don't understand hdmi either so it might be a necessary evil. 

 

Allways an issue with specifications which can do a lot if things, it's either confusing (complicated) or too simplified...which causes different confusion when things are assumed to work when it is not that simple....case and point hdmi "versions" which simplified branding too much. 

HDMI doesn't have versions implying features either and they say to explicitely mention the features you implement (VRR, ALLM etc.) instead of errounously using the specification version. HDMI's marketing labels are the "Ultra High Speed HDMI Cable" things, for example.

3 hours ago, infigo said:

But current usb 4 usb 20 gigabits is still usb 4. 

Usb4 CAN use thunderbolt for more speed and driect PCIe connection...but don't have to, can still use usb as protcoll  which is identical to usb 3.2 gen2x2....but it is usb4. 

So we are already passed that notion. 

 

USB4 is based on Thunderbolt 3. It can tunnel (as they call it) the USB 3.2 protocol and use the Gen 2x2 transfer mode if requested. Again the devices can negotiate this among themselves, so all the consumer needs to know is that the it can do 20 Gbit/s.

[infigo]

14 hours ago, tikker said:

The entire point is that you as a consumer generally don't need to know whether it is electrically dual or single-lane. All you need to know is that it can do 20 Gbit/s and the device will internally figure out how to do that. USB is made to be backwards compatible, so a newer device should be able to accomodate an older device or transfer mode easily. The example you give is also one of the main points of USB4, so I think that should also be a non-issue, since that is part of the design. 

 

Do you have an example where you explicitely had to know what the supported transfer modes were? The feature is X Gbit/s transfer speeds, not a specific transfer mode. The spec states that any port supporting USB 3.2 connected to a device supporting USB 3.2 should support the USB 3.2 SuperSpeed and SuperSpeedPlus modes, which I think implies all the transfer modes should be supported anyway. That would make this a non-issue.

USB4 is based on Thunderbolt 3. It can tunnel (as they call it) the USB 3.2 protocol and use the Gen 2x2 transfer mode if requested. Again the devices can negotiate this among themselves, so all the consumer needs to know is that the it can do 20 Gbit/s.

 

If, in your example, the cable/device can sustain 20 Gbit/s transfer in addition to other devices that means it can do more than 20 Gbit/s and thus  it should likely be a 40 Gbit/s cable, for example, in the first place. That would make it irrelevant again, to the consumer, how that 20 Gbit/s is achieved.

Lets see if I can answer this in a way without loosing the red thread. Usb and thunderbolt is both included in usb 4 but devices could use either or. The 40/80 gigabit of new usb4 specification (whatever they call it) is to my understanding usb based not thunderbolt based. So a 40 gigabit may be thunderbolt or usb as a protocoll. 

 

It may not matter to consumers but as HDMI has showed that the branding consumers has shown (ie the version numbers) aren't representative as you say HDMI is a collection of features, where bandwidth is one of them. Few remember intel hd 2000/3000 (if I remember right) not supporting 4k60hz (or was it like 21:9 1440p 60hz or 90 hz, was too long for me to remember) over hdmi as they were just shy of the required bandwidth for it....even though hdmi specification 1.4b introduced bandwidth for it 1-2 years prior. People get stuck on the branding name. 

 

USB-IF might be choose a better middleground, actually advertising the speed and not the specification but I fear it will lead to situations where MST hubs won't work as advertised as it is requireing single lane usb (gen1,gen2 or gen3 speed). I want to say fuck it to displaylink docks but as thunderbolt is too expensinve of and implementation (may change), mst docks is what I believe is finally at a point soon to be a one connector solution without hassle. But it is dependent on new bandwidth of dp2.0 (to support dual 4k on a dp single pair with mst) and having good usb branding which people won't misunderstand. 

 

A mst displayport solution can never work in a situation where usb is running in genYx2 mode. But it also requires good feature communication from device manifacturer...which historically have been useless, both for usb-c and hdmi 

[tikker]

1 hour ago, infigo said:

Lets see if I can answer this in a way without loosing the red thread. Usb and thunderbolt is both included in usb 4 but devices could use either or. The 40/80 gigabit of new usb4 specification (whatever they call it) is to my understanding usb based not thunderbolt based. So a 40 gigabit may be thunderbolt or usb as a protocoll.

USB4 is Thunderbolt based: https://arstechnica.com/gadgets/2019/03/thunderbolt-3-becomes-usb4-as-intels-interconnect-goes-royalty-free/

It's a different protocol from USB 3.2, but it can can accomodate the previous modes to be used. To my knowledge the integration of TB3 is the reason behind the speed boost to 40 Gbit/s (all TB3 cables should still be compatible).

1 hour ago, infigo said:

t may not matter to consumers but as HDMI has showed that the branding consumers has shown (ie the version numbers) aren't representative as you say HDMI is a collection of features, where bandwidth is one of them. Few remember intel hd 2000/3000 (if I remember right) not supporting 4k60hz (or was it like 21:9 1440p 60hz or 90 hz, was too long for me to remember) over hdmi as they were just shy of the required bandwidth for it....even though hdmi specification 1.4b introduced bandwidth for it 1-2 years prior. People get stuck on the branding name. 

From what I remember the biggest HDMI 1.4 debacle was the 1080p@120Hz stuff. The maximum bandwidth in the 1.4b spec is 8.16 Gbit/s, which couldn't support 4k60, but it's a good example. As far as I know the 1.4b version number also referred just to the HDMI specification. While the spec can introduce a capability, it is up to the manufacturer to implement that and to honestly market it being capable of that resolution and refresh rate. Manufacturers didn't really do that and just used the version number, leading to consumers wrongly thinking that since it's in the spec it must be present on the device.

1 hour ago, infigo said:

USB-IF might be choose a better middleground, actually advertising the speed and not the specification

That is what they are doing though. It's manufacturers that use version numbers and transfer modes as some have shown in earlier examples. The marketing names for USB4 are "USB4 20 Gbit/s" and "USB4 40 Gbit/s". The specification (of which Version 2 was announced per this thread) is USB4 and Gen 2x2, for example, is a transfer mode and not a specification (sub)version. That's why I am largely blaming manufacturers, because even for USB 3.2 the official marketing name is "SuperSpeed USB 20 Gbit/s", for example.

 

Internally the need a way to distinguish how the transfer is achieved exactly, hence why the Gen 1, 2, 3 and YxZ transfer modes exist. They are not really versioning, nor marketing material however.

2 hours ago, infigo said:

but I fear it will lead to situations where MST hubs won't work as advertised as it is requireing single lane usb (gen1,gen2 or gen3 speed). I want to say fuck it to displaylink docks but as thunderbolt is too expensinve of and implementation (may change), mst docks is what I believe is finally at a point soon to be a one connector solution without hassle. But it is dependent on new bandwidth of dp2.0 (to support dual 4k on a dp single pair with mst) and having good usb branding which people won't misunderstand. 

I don't think this will happen as I don't think you can require a specific Gen X YxZ implementation. You have to implement a protocol, e.g. USB 3.2, at a certain speed which will come with certain transfer modes. I therefore stick with the opinion that it is up to the manufacturers to clearly advertise what their product is capable of (and do also support de opinion that USB-IF could be more proactive about ensuring that).

2 hours ago, infigo said:

A mst displayport solution can never work in a situation where usb is running in genYx2 mode. But it also requires good feature communication from device manifacturer...which historically have been useless, both for usb-c and hdmi 

Are there any tests backing that up? AFAIK that will be implemented as one of the alternate modes, which is part of the USB-C specification (not USB 3.2 or USB4) and thus has rules on how it should be implemented as well.

[Shzzit]

So I could run one usb wire to a hub and connect all my stuff too that with no added latency or input lag? 

[Eigenvektor]

4 hours ago, Shzzit said:

So I could run one usb wire to a hub and connect all my stuff too that with no added latency or input lag? 

Increased bandwidth does not decrease latency. It will still take the same time to transmit signals, you can simply send more stuff at the same time.