• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

The USB-IF Announces New Logos, Kills Off SuperSpeed Branding

IF its not SuperSpeed...

:pimp:



Whats in a name? color coding is better. :p
 
but what about the 80 Gbps cables? ;)
End of next year. It's early days. We barely have any 40 Gbps cables. They need to validate that both DP 2.0 Alt Mode and tunnelled work well on all new USB hardware. Currently, DP 1.4 is supported.
They also need to develop commercial switch chip, similar to TB4. That will need to be PCIe 5.0 x4 connection providing up to 128 Gbps, so that two ports can use the bandwidth.

Gosh it's as confusing as before. Just with new logos.
This is much better than it used to be. What is confusing you?

I honestly hate USB at this point and dislike that it is replacing so many other ports. I want my dedicated ports!
Which dedicated ports? Majority of new motherboards have HDMI and DP ports. They are adding USB-C now too, that tunnels DP from iGPU.

So USB has two different 20 Gbps modes… there is USB 3.2x2 = 20 gbps, but with no pcie tunneling. USB4 also has a 20 gbps mode that uses different signaling than USB 3, and IIRC, it can tunnel pcie at 20 gbps.

so when we see USB20Gbps in new equipment, which version is it, usb3.2x2 with no pcie tunneling, or is it usb4? Will we know?
Good question. USB20 operates over USB-C port. I have not seen one single new motherboard with USB-A at 20 Gbps. So, that's out of the way.
If you see on USB-C port USB20 label only (currently SS20 on most boards or 20G on Gigabyte), it's going to be USB data only.
If you see on USB-C port USB4 or USB10 with DP logo, this is where DP and PCIe tunnelling are implemented.

We will not see PCIe letters on rear I/O, but Microsoft mandates now every vendor to support PCIe tunnelling on every USB4 port, 20 or 40 Gbps. Most vendors will go with 40 Gbps ports, as DP 1.4 tunnelling needs to be supported at 32 Gbps, just like Thunderbolt 4 does. I have not seen one single USB4 solution with 20 Gbps implementation, as this would mean that even DP 1.2 cannot be supported at 21 Gbps. Hope this clarify matters.

Below is what Microsoft says about it.
USB4 PCIe Tunneling must on Windows.png


The whole PCIe tunneling feature is all but abandoned. example: WIthout known exceptions, all the X670E boards are using Intel Thunderbolt 4 controllers for their USB 4, but have no advertised support for PCIe tunneling
No. Thunderbolt 4 supports PCIe protocol by deafult. It's in the spec. External GPUs work on Thunderbolt 4 ports and transmit PCIe data in both directions.
Since the one major differentiating feature is effectvely depreciated before even hitting the market, the difference between 20gbps usb3 and 20gbps usb4 is a non-issue for the consumer-facing marketing/brandng.
No. This is a mute point. PCIe tunnelling is mandatory on all USB4 ports on Microsoft devices. Read XDA pages and my post above. Whoever installs USB4 port, it will need to be 40 Gbps in order to support DP 1.4 and PCIe tunnelling. 20 Gbps USB4 port cannot support even DP 1.2 at 21 Gbps.
 
Last edited:
How is USB 4.x so hard to do?
Fully agree with you!
Simple is the best.
For example.
Usb 4.5
Usb 4.10
Usb 4.20
Usb 4.40
And etc
 
(Looking for USB 3.2 nvme enclosures, i stumbled upon many a content-creator using USB 3.2 external drives for storing raw footage. Years gone by, this would've been Firewire and Thunderbolt) Details matter, but for consumer electronics, manufactureres typically prefer to disclose as few of technical details as possible.
NVMe enclosures can work in several ways, depending on the chip inside and whether it's directly connected to a host or via hub. It's complex. Here we go. The most important question is what chip is inside of NVMe drive enclosure and what chip is in the host system.

What do vendors install in NVMe enclosures:
1. USB 3.2 chip 10 or 20 Gbps for direct host-peripheral connection. Both need to support the same speed. The same for tunnelled, but 20 Gbps is optional.
2. Thunderbolt chip at x2 or x4 PCIe lanes (many vendors install older TB3 6000 chip with x2 PCIe 3.0). USB 10 Gbps is mandatory, can be tunnelled too.
A. If the host has TB, PCIe connection at 16 Gbps or 32 Gbps can be established (the fastest transfer)
B. If the host does not have TB, but only USB-C port, USB at 10 Gbps should work
C. if the host has USB4 port, it shoud work like TB
3. USB4 chip is more complex. Peripheral can have either 20 Gbps or 40 Gbps chip. In tunnelled mode, USB 10 Gbps is mandatory and PCIe function is optional.
A - if the host has the same USB4 chip as the peripheral, 20 or 40 Gbps will work; PCIe tunnelling in mandatory, so look out if PCIe is supported on drive too
B - if the host has TB, USB will work at 10 Gbps, but if the peripheral supports PCIe funtion, NVMe drive can connect at 16 or 32 Gbps

Ah, I was worried there would be no USB rebranding stupidity this week. USB-IF is full of drunk or drugged monkeys, I see no other explanation.
What they did is much better more simplified now. Don't you realize it?

but the problem is one version of usb20gbps (usb3.2 genx2) is usb mode only with no pcie tunneling and the other (usb4) may support pcie tunneling. The label “usb 20 gbps” is ambiguous and you don’t know if the “usb 20 gbps port”supports pcie tunelling or not.
This has been explained. The host must support PCIe tunnelling due to Microsoft requirements for Windows host devices. For peripherals, it remains optional.
So if you connect an alpine ridge device to a “usb 20gbps” port, it may connect or it may not, depending on if the “usb 20 gbps” port supports pcie tunneling and tb3 signaling. Alpine ridge does not have a usb fallback mode. So if the “usb 20 gbps” port is really a usb3.2 gen2x2 port, the alpine ridge device will fail to connect. But if the “usb 20 gbps” port can tunnel pcie and support to signaling, then alpine ridge devices will be able to connect. Titan ridge has a usb-fallback mode and doesn’t have this problem.
Have you tried those connections? I explained this too above.
- SS20 or 20G labelled ports are simple USB ports and will accept Alpine Ridge devices as USB devices at 10 Gbps
- 40 DP or 10G DP or SS10 DP labelled ports should support DP and PCIe tunelling, as those are USB4 ports. Alpine Ridge peripheral should connect to those ports at 16 Gbps (JHL6240) or at 32 Gbps (JHL6540 and JHL6340) for PCIe data.

The most secure option on the host is 40 DP, as this is JHL8540 Thunderbolt 4 chip and we know how it works. Asrock and Asus are doing a good job here.
Bizarrely, several new motherboards with this chip do not have Intel's "thunder" logo, meaning that those chips have de facto became USB4 chips. We have not reached uniform labelling, as yet...
Rear io.JPG


However, some motherboard vendors that did not install JHL8540 as their 'USB4' solution, may skip PCIe tunnelling by labelling USB ports as SS10 DP, and simply omit PCIe by providing DP from onboard graphics alongside USB10 (Gigabyte and MSI). This is to be checked once someone tests those ports.

It's almost like we should have just kept Thunderbolt Thunderbolt and DisplayPort DisplayPort instead of cramming 101 things into USB...
No. It is Intel that has "crammed" Thunderbolt PHY and features into USB4 specification, to harmomize the interface over USB-C. They even told USB agency that USB should never be as good in the spec as TB is. I hope USB IF showed them a middle finger on that. Both bodies are now pursuing the next level, which is 80 Gbps with PCIe 5.0 x4 and DP 2.0. It's a race.
 
Last edited:
It's still a fustercluck
I personally liked the simplicity of Thunderbolt 3 and 4. It spells out, in plain language, what you're getting. (The only minor exception is if your device supports one or two DP streams for things like LG's 5k TB3 displays. One stream limits you to 4k like the ports on my HP Spectre.)

USB is one port that might* do everything. TB3/4 is one port that does do everything. To me, that's a pretty big distinction.
*: YMMV.
 
At this rate it'll be like shopping for tires...

USB4 40/w240
 
I personally liked the simplicity of Thunderbolt 3 and 4. It spells out, in plain language, what you're getting. (The only minor exception is if your device supports one or two DP streams for things like LG's 5k TB3 displays. One stream limits you to 4k like the ports on my HP Spectre.)

USB is one port that might* do everything. TB3/4 is one port that does do everything. To me, that's a pretty big distinction.
*: YMMV.
Actually, the USB4 implementations so far, when it comes to host controllers, are full-fat, full-feature host controllers, so they offer better features than Thunderbolt 4, as they can go up to the full 40 Gbps for data transfers, whereas Thunderbolt 4 is limited to to 32 Gbps today, due to the Intel chips being PCIe 3.0 and not PCIe 4.0. Admittedly ASMedia is late to the game, which is why we're seeing Intel chips being repurposed and re-named USB4. USB4 can also do 20 and 40 Gbps over the USB standard, where Thunderbolt chips from Intel are limited to 10 Gbps.
Yes, Thunderbolt 4 has some features that aren't supported by USB4, but only if the controller is built into an Intel based system, as some features are Intel specific.
In fact, despite being a key member of the USB-IF, Intel is trying hard to make USB4 look inferior to Thunderbolt 4, when in fact, USB4 is often at parity on the host side, sometimes exceeding Thunderbolt 4, for example when combined with USB PD.
IS USB4 perfect? Far from it, which I why I wrote this a few months ago, but it's also by no means crippled, as Intel is trying to make it out to be. Some devices will be limited, but most of those devices aren't likely to need a full implemenation of USB4 either.

At this rate it'll be like shopping for tires...

USB4 40/w240
At least you went for the top of the range.
 
Might as well, then you'll know the cable will be backwards compatible with the other specs (at least it should). Guess we'll find out when LTT Labs and others test them.
 
Death by a million cuts specs. The one thing Thunderbolt didn't get wrong.

Insert too many standards +1 xkcd panel :D

Also, can't wait for Cheap under-spec cables being sold as 40Gbps 240W cables, then melting/catching fire/killing devices due to being USB 3 Gen 1 cables with 28-32awg throughout.

USB cables now require an identifier chip that should prevent that, it won't prevent false advertising but unless there's really gross negligence should prevent fires or other hazards. The increased cost alone should keep the more sleezy players out of the market.

Good question. USB20 operates over USB-C port. I have not seen one single new motherboard with USB-A at 20 Gbps. So, that's out of the way.

USB-A can't do 20gbps ever, that requires 2 differential pairs which USB-A doesn't have. It was why the "2x2" stupid branding became a thing (it used the 2 pairs available on USB-C)

NVMe enclosures can work in several ways, depending on the chip inside and whether it's directly connected to a host or via hub:

1. single USB 3.2 chip 10 or 20 Gbps for direct host-peripheral connection. Both need to support the same speed. The same for tunnelled, but 20 Gbps is optional.
2. Thunderbolt chip at x2 or x4 PCIe lanes (many vendors install older TB3 6000 chip with x2 PCIe 3.0). USB 10 Gbps is mandatory, can be tunnelled too.
A. If the host has TB, PCIe connection at 16 Gbps or 32 Gbps can be established (the fastest transfer)
B. If the host does not have TB, but only USB-C port, USB at 10 Gbps should work
C. if the host has USB4 port, it shoud work like TB
3. USB4 is more complex here. Peripheral can have either 20 Gbps or 40 Gbps chip. In tunnelled mode, USB 10 Gbps is mandatory and PCIe function is optional.
A - if the host has the same USB4 chip as the peripheral, 20 or 40 Gbps will work; PCIe tunnelling in mandatory, so look out if PCIe is supported on drive too
B - if the host has TB, USB will work at 10 Gbps, but if the peripheral supports PCIe funtion, NVMe drive can connect at 16 or 32 Gbps

TLDR: that's why this sucks, I just want to plug my device and for it to work fast ;)
It's the reason iFruit stuff has a supposed "good user experience", you don't need a degree to figure out what will work or not, you also don't have nearly as many options and are often limited to the aproved options, but they work and you don't need to know how or why.

No. It is Intel that has "crammed" Thunderbolt PHY and features into USB4 specification, to harmomize the interface over USB-C. They even told USB agency that USB should never be as good in the spec as TB is. I hope USB IF showed them a middle finger on that. Both bodies are now pursuing the next level, which is 80 Gbps with PCIe 5.0 x4 and DP 2.0. It's a race.

Given the optional stuff in USB4 and how Thunderbolt 4 is clear and has almost everything mandatory to meet certification, doesn't seem to have been any middle figure.

I personally liked the simplicity of Thunderbolt 3 and 4. It spells out, in plain language, what you're getting. (The only minor exception is if your device supports one or two DP streams for things like LG's 5k TB3 displays. One stream limits you to 4k like the ports on my HP Spectre.)

Thunderbolt 3 had the problem of being 2 or 4 lanes wide which allowed sleezy manufacturers to trick users, Thunderbolt 4 solved that and made most of the optional stuff required and always 4 lanes.
 
Killing of the "Super Speed" branding?
Really?
in 2022 "SS" is not sounding good? 'bout time :)
Aside of joking
I would drop the stupid Gbps and use GB/s numbering instead, but it's we can hope for better I guess.
 
End of next year. It's early days. We barely have any 40 Gbps cables. They need to validate that both DP 2.0 Alt Mode and tunnelled work well on all new USB hardware. Currently, DP 1.4 is supported.
They also need to develop commercial switch chip, similar to TB4. That will need to be PCIe 5.0 x4 connection providing up to 128 Gbps, so that two ports can use the bandwidth.


This is much better than it used to be. What is confusing you?


Which dedicated ports? Majority of new motherboards have HDMI and DP ports. They are adding USB-C now too, that tunnels DP from iGPU.


Good question. USB20 operates over USB-C port. I have not seen one single new motherboard with USB-A at 20 Gbps. So, that's out of the way.
If you see on USB-C port USB20 label only (currently SS20 on most boards or 20G on Gigabyte), it's going to be USB data only.
If you see on USB-C port USB4 or USB10 with DP logo, this is where DP and PCIe tunnelling are implemented.

We will not see PCIe letters on rear I/O, but Microsoft mandates now every vendor to support PCIe tunnelling on every USB4 port, 20 or 40 Gbps. Most vendors will go with 40 Gbps ports, as DP 1.4 tunnelling needs to be supported at 32 Gbps, just like Thunderbolt 4 does. I have not seen one single USB4 solution with 20 Gbps implementation, as this would mean that even DP 1.2 cannot be supported at 21 Gbps. Hope this clarify matters.

Below is what Microsoft says about it.
View attachment 263969


No. Thunderbolt 4 supports PCIe protocol by deafult. It's in the spec. External GPUs work on Thunderbolt 4 ports and transmit PCIe data in both directions.

No. This is a moot point. PCIe tunnelling is mandatory on all USB4 ports on Microsoft devices. Read XDA pages and my post above. Whoever installs USB4 port, it will need to be 40 Gbps in order to support DP 1.4 and PCIe tunnelling. 20 Gbps USB4 port cannot support even DP 1.2 at 21 Gbps.
I think the usb-if’s attempt at simplicity will make things even more confusing. USB4 and thunderbolt are tunneled protocols. USB4 has optional modes that are not mandatory. Just because a cable or a port says 20 Gbps or 40 Gbps doesn’t mean that all the optional features like pcie tunneling, DisplayPort 1.4, etc have been enabled.

I like that Microsoft made pcie tunneling mandatory for usb4, but that is a Microsoft decision, not usb-if. I think the labeling of a cable and/or port should have check marks for all the protocols supported. Usb 20 Gbps/40Gbps. Then check marks for: dp1.4, pcie, etc.
 
Thunderbolt 3 had the problem of being 2 or 4 lanes wide which allowed sleezy manufacturers to trick users, Thunderbolt 4 solved that and made most of the optional stuff required and always 4 lanes.
Which is still far less fragmented than USB. :p
 
I think the usb-if’s attempt at simplicity will make things even more confusing. USB4 and thunderbolt are tunneled protocols. USB4 has optional modes that are not mandatory. Just because a cable or a port says 20 Gbps or 40 Gbps doesn’t mean that all the optional features like pcie tunneling, DisplayPort 1.4, etc have been enabled.

I like that Microsoft made pcie tunneling mandatory for usb4, but that is a Microsoft decision, not usb-if. I think the labeling of a cable and/or port should have check marks for all the protocols supported. Usb 20 Gbps/40Gbps. Then check marks for: dp1.4, pcie, etc.
Why would you have to look for that? Data is data.
If a cable can do 40 Gbps speeds, then it can handle everything else that USB4 would be able to do.

This is the reason why it's such a mess:


Too many cooks in the kitchen, design by committee, etc. :D
Many of those cooks wanted a better spec.
Intel said no.
 
It doesn't matter for the cables, but it does for the ports.
On what exactly? I mean, if a device doesn't support something, then it doesn't matter what the cable or host supports.
 
Why would you have to look for that? Data is data.
If a cable can do 40 Gbps speeds, then it can handle everything else that USB4 would be able to do.


Many of those cooks wanted a better spec.
Intel said no.
Data is not data bro. Usb3.2 gen2x2 can do 20 Gbps but does not tunnel pcie. USB4 has a 20 gbps mode, and can tunnel pcie. So there are two usb 20 Gbps modes, one with pcie tunneling and one with no pcie tunneling.

So if you just look at a usb label that says 20 Gbps… how do you know with any confidence pcie tunneling enabled? What about display port ?
 
USB4 implementations so far, when it comes to host controllers, are full-fat, full-feature host controllers, so they offer better features than Thunderbolt 4, as they can go up to the full 40 Gbps for data transfers, whereas Thunderbolt 4 is limited to to 32 Gbps today, due to the Intel chips being PCIe 3.0 and not PCIe 4.0.
This is yet to be confirmed when ASM4242 us tested. ASM4242 connnects to either CPU or chipset lanes at PCIe 4.0 x4, so 64 Gbps. If it's a one usb-c port solution, then yes, it gets up to 40 Gbps of PCIe data. But this solution is wasteful of almost two PCIe lanes.

If it's two usb-c ports solution, it is not clear as yet how those ports are fed with PCIe lanes. Does each port have PCIe 4.0 x2 assigned, which would mean each port gets up to 32 Gbps, or is there PCIe 4.0 switch, like in TB4, that assigns bandwidth dinamically, depending on traffic on each port? If the latter, one port can get up to 40 Gbps, assuming full traffic, and another one up to remaining 24 Gbps. I hope Asmedia has installed PCIe switch for dynamic traffic.
 
Data is not data bro. Usb3.2 gen2x2 can do 20 Gbps but does not tunnel pcie. USB4 has a 20 gbps mode, and can tunnel pcie. So there are two usb 20 Gbps modes, one with pcie tunneling and one with no pcie tunneling.

So if you just look at a usb label that says 20 Gbps… how do you know with any confidence pcie tunneling enabled? What about display port ?
Did you read my USB4 primer article?

I guess not, based on your comments.
If it says 20 Gbps, that's all you need to know. Tunneling will work, regardless of what you're sending in the tunnel, that's why tunnelling is so great.

This is yet to be confirmed when ASM4242 us tested. ASM4242 connnects to either CPU or chipset lanes at PCIe 4.0 x4, so 64 Gbps. If it's a one usb-c port solution, then yes, it gets up to 40 Gbps of PCIe data. But this solution is wasteful of almost two PCIe lanes.

If it's two usb-c ports solution, it is not clear as yet how those ports are fed with PCIe lanes. Does each port have PCIe 4.0 x2 assigned, which would mean each port gets up to 32 Gbps, or is there PCIe 4.0 switch, like in TB4, that assigns bandwidth dinamically, depending on traffic on each port? If the latter, one port can get up to 40 Gbps, assuming full traffic, and another one up to remaining 24 Gbps. I hope Asmedia has installed PCIe switch for dynamic traffic.
See article above. I visited ASMedia and talked to them about this. One port can do 40 Gbps while the other can do 20 Gbps simultaneously.
 
In fact, despite being a key member of the USB-IF, Intel is trying hard to make USB4 look inferior to Thunderbolt 4, when in fact, USB4 is often at parity on the host side, sometimes exceeding Thunderbolt 4, for example when combined with USB PD.
So true. We shall see whether Thunderbolt 5 launches earlier than USB4 V2. The points of convergence and divergence:
- both will have 80 Gbps bidirectional ports and cables
- both will have DP 2.0 Alt More for 80 Gbps traffic to monitors
- Intel is under pressure to provide the latest PD 3.1 on TB5 and step up the game to 240W, to feed monitors and more needy laptops
- Intel also needs to increase USB throughput up from 10 Gbps to 40 Gbps, tunnelled or not
- USB IF should enable networking at 10 GbE, Intel has an edge here
- PCIe PHY is a bit of a mistery. Both organisations should go for PCIe 5.0 x4 to the host, to get access to 128 Gbps, so that each usb-c port could operate at Gen4 x4 to use 64 Gbps, or dynamically use up to 80 Gbps on one port. This would be ideal for AMD 7000 CPUs, as those already provide PCIe 5.0 x4 for such link, in addition to Gen5 x16 for GPU and Gen5 x4 for NVMe drive. All this directly on CPU! Intel's CPUs in Meteor Lake 14th will need to add another eight Gen5 lanes to match AMD CPUs in this regard.

Many of those cooks wanted a better spec.
Intel said no.
Why AMD is not the part?

So if you just look at a usb label that says 20 Gbps… how do you know with any confidence pcie tunneling enabled? What about display port ?
If you see on USB-C port USB20 label only (currently SS20 on most boards or 20G on Gigabyte), it's going to be USB data only.
If you see on USB-C port USB4 or USB10 with DP logo, this is where DP and PCIe tunnelling are implemented.

We will not see PCIe letters on rear I/O, but Microsoft mandates now every vendor to support PCIe tunnelling on every USB4 port, 20 or 40 Gbps. Most vendors will go with 40 Gbps ports, as DP 1.4 tunnelling needs to be supported at 32 Gbps, just like Thunderbolt 4 does. I have not seen one single USB4 solution with 20 Gbps implementation, as this would mean that even DP 1.2 cannot be supported at 21 Gbps. Hope this clarify matters.

Thunderbolt 3 had the problem of being 2 or 4 lanes wide which allowed sleezy manufacturers to trick users, Thunderbolt 4 solved that and made most of the optional stuff required and always 4 lanes.
True that. Asrock Z390 ITX board has Apline Ridge chip with PCIe 3.0 x2 lanes. They never mentioned it anywhere. They said: "Thunderbolt 3 40 Gbps"...

See article above. I visited ASMedia and talked to them about this. One port can do 40 Gbps while the other can do 20 Gbps simultaneously.
So, the host chip is not ready yet, right? I have checked all AMD X670 and B650 motherboards and none features ASM4242 chip, only JHL7540.
It might appear in the second wave of motherboards next year.
 
So, the host chip is not ready yet, right? I have checked all AMD X670 and B650 motherboards and none features ASM4242 chip, only JHL7540.
It might appear in the second wave of motherboards next year.
ASMedia is late, so even though it was meant to be on the first wave of boards, it's just not happening.
It's a real shame, but I guess part of the holdup is the certification as well, I mean, even AMD shipped their first 6000-series notebook SoCs without USB4 support, it came later via a UEFI and driver update.
As for the ASM4242, I guess board makers could've gone down the same route, had the chip been in mass production. I wasn't told when it would be in mass production, as it was confidential.

Why AMD is not the part?
It was Intel that "donated" TB3 to the USB-IF, so I guess they got to decide the USB4 base specs somehow.
 
It was Intel that "donated" TB3 to the USB-IF, so I guess they got to decide the USB4 base specs somehow.
I meant as founding or current members of the group.
Anyway, since Intel has released GPUs with DisplayPort 2.0 at 40 Gbps, do you have any hint about DP 2.0 monitors? Even TFT is not able to gather any meaningful information about it.

It sounds bizarre that nobody publishes anything substantial about it. Even since VESA's rep had an interview with Tom's Hardware in Janaury 2021!, he said "end of 2021". It's end of 2022 and zero news. I am confused.

The only news I have heard about it was when AMD officially certified DP 2.0 port on Rembrandt laptop in San Jose during display symposium. There was a mention of "reference monitor" with MediaTek DP 2.0 chip inside.

It all sounds secretive, hush, hush, unclear, muddy. Are there any announcements in the pipeline that you are aware of? Do you guys plan an article about it? Perhaps a follow up interview with the same VESA rep Craig W.?

Also, there are no leaks whatsoever. Are any DP 2.0 monitors going to be announced alongside RDNA3 launch? It's usually then and during CES that we get some info.
 
Death by a million cuts specs. The one thing Thunderbolt didn't get wrong.
It also is not that easy ;)
E.g. TB3 cable or TB3 port can be full speed (40 Gbps), but it can be also limited to half speed (20 Gbps).
And usually it isn't clearly stated. Especially with notebooks ports.
 
I do look forward to 40Gb+ PCI-E tunneling for future devices

It should absolutely shake up the laptop world, imagine when any laptop can be docked to a PCI-E 4.0/5.0 x4 external GPU?
 
Back
Top