beelzebozo
Super Moderator
- Joined
- Mar 31, 2010
- Messages
- 3,723
- Motherboard
- Gigabyte z690-AERO-G-DDR4
- CPU
- i7-13700K
- Graphics
- RX 6900 XT
- Mac
- Classic Mac
- Mobile Phone
USB 4 Version 2 - 80gbps
- Thread starter beelzebozo
- Start date
- Status
- Joined
- Oct 4, 2018
- Messages
- 1,489
- Motherboard
- Gigabyte C246-WU4
- CPU
- E-2278G
- Graphics
- WX7100
- Mac
Useful summary on USB 3.x, USB4, TB3 and TB4.
arstechnica.com
Breaking down how USB4 goes where no USB standard has gone before
USB4 vs. Thunderbolt 4—and everything else to know about the newest USB standard.
arstechnica.com
pastrychef
Moderator
- Joined
- May 29, 2013
- Messages
- 19,458
- Motherboard
- Mac Studio - Mac13,1
- CPU
- M1 Max
- Graphics
- 32 Core
- Mac
- Classic Mac
- Mobile Phone
Where does this put Thunderbolt? Will it die the same way Firewire died?
- Joined
- Oct 4, 2018
- Messages
- 1,489
- Motherboard
- Gigabyte C246-WU4
- CPU
- E-2278G
- Graphics
- WX7100
- Mac
That is—and possibly always was—a distinct possibility. Or Intel will come up with some "Thunderbolt 5" which incorporates most or all of "USB4 v.2" (or whatever it is called), and possibly some more. (And, at the very least, gets the name right and simple!)
From the comments on news sites, many PC users weren't even aware there was something beyond USB 3.2. And the present piece of news is an announcement that the specification for USB4 version 2 will be published in November, with controllers and cables to follow even later. There is still time for Thunderbolt to evolve.
UtterDisbelief
Moderator
- Joined
- Feb 13, 2012
- Messages
- 9,622
- Motherboard
- Gigabyte B760 Gaming X AX
- CPU
- i5-14600K
- Graphics
- RX 560
- Mac
- Classic Mac
- Mobile Phone
Can you imagine the confusion the different cables will cause? Not to mention the high cost. Right now USB-C 3.2 Gen 2 fully connected cables are not cheap. What will a 240W cable be? Will it need to be very short if passive and active if long? Will it glow red-hot in use?!
It's great that the standards are being pushed to greater heights, but to be commonplace the complexity needs rationalising. And in a way maybe a new port design would be "kinder" to the end-user because you would know where you stood.
It's great that the standards are being pushed to greater heights, but to be commonplace the complexity needs rationalising. And in a way maybe a new port design would be "kinder" to the end-user because you would know where you stood.
pastrychef
Moderator
- Joined
- May 29, 2013
- Messages
- 19,458
- Motherboard
- Mac Studio - Mac13,1
- CPU
- M1 Max
- Graphics
- 32 Core
- Mac
- Classic Mac
- Mobile Phone
In my opinion, the lack of any significant advancements to Thunderbolt 4 was a strong tell. USB4 brought the two standards pretty much to performance parity. Now, USB4 v2 has surpassed Thunderbolt.
I strongly believe that Thunderbolt's fate will end up being the same as Firewire's. The lack of widespread adaption in PCs is its death knell. Even though Thunderbolt is found on all current Macs, Thunderbolt accessories are not used as often due to the price premiums. Now that Thunderbolt's performance has been surpassed, there's no reason at all to pay those price premiums.
Last edited:
- Joined
- Apr 12, 2021
- Messages
- 901
- Motherboard
- Asus z590 ROG Maximus XIII Hero
- CPU
- i9-11900K
- Graphics
- RX 6600 XT
- Mac
- Classic Mac
- Mobile Phone
I didn't read these articles carefully.
Was there mention of throwing some key aspect of TB away?
The way I read it, there's no reason at all for concern of TB to die: it already got completely merged into USB4. And USB has been very backwards compatible.
TB has never been at odds with USB. And the connectors have already been harmonized. —I find it a bit strange there aren't locking versions of the connectors, are there? The connector is sort of magic to me because it's such an obviously mechanical thing yet made from parts almost too small to see with naked eye. I'm curious about the machines that manufacture the connectors. I don't get same feeling of uncanny from chips. But I also never tell amazed that electrons do their freaky stuff.
WRT to marriage of TB add USB: The formal distinction of TB4 to USB4 is that TB is the most rigorous application instance of USB, mandating top capabilities, whereas USB4 lets vendors choose a wide level of support over a wide range. You could maybe try to argue too wide, but when have PC nerds ever complained about too many options Those are the selling points! I will be completely astounded if there's a backlash any PC user defects to Apple because they finally realized they need a more restrictive and tightly curated applications, but if I ever find such a specimen it makes sense if it's a hackintosher who cracked over how much complexity he'll put up with.
USB4 V2 continues to evolve TB and further builds out protocol.
To the extent that TB features require specifically rated cables, there's a minor headache of knowing what's needed. The practical limits are due to extremely well understood physics of which there's no danger of being unlearned to the degree that you expect the chips to continue to get made. Wire heating is a direct function of current can be mediated with higher voltages. If they're doing 240W at 5V that's 50 amps! Which seems like a lot. But cable resistance can't be more that a tiny fraction of an ohm, and the connector already needs to have fantastic integrity for 50 gigahertz signaling, which is the real mystery.
The only thing that feels bad to me about this tech is the branding is dumb. It could use a bit of Steve touch.
What it looks like is that with USB4 V2 we stand at threshold of a general-purpose multicomputer interconnect technology at a bargain-basement consumer application price-point. If there's a remote DMA capability lurking in the specs then doors might open to a very interesting future, per novel arrangements of chiplets.
We again stand on shoulders of giants. To me it feels like a total heyday!
The attending morbidity on this point is odd. As if there's a lost ability to dream about new ways to personalize the tech. Maybe such dreams have dissipated because the physics are so close to magic. OTOH, we've got DALL-E and Stable Diffusion to dream for us. Whee!
Let's see I just plug my tonymacx86 bat cave post generator dongle into this usb port here, and... Doh! I didn't map that port as type 65535, garrr!
Was there mention of throwing some key aspect of TB away?
The way I read it, there's no reason at all for concern of TB to die: it already got completely merged into USB4. And USB has been very backwards compatible.
TB has never been at odds with USB. And the connectors have already been harmonized. —I find it a bit strange there aren't locking versions of the connectors, are there? The connector is sort of magic to me because it's such an obviously mechanical thing yet made from parts almost too small to see with naked eye. I'm curious about the machines that manufacture the connectors. I don't get same feeling of uncanny from chips. But I also never tell amazed that electrons do their freaky stuff.
WRT to marriage of TB add USB: The formal distinction of TB4 to USB4 is that TB is the most rigorous application instance of USB, mandating top capabilities, whereas USB4 lets vendors choose a wide level of support over a wide range. You could maybe try to argue too wide, but when have PC nerds ever complained about too many options Those are the selling points! I will be completely astounded if there's a backlash any PC user defects to Apple because they finally realized they need a more restrictive and tightly curated applications, but if I ever find such a specimen it makes sense if it's a hackintosher who cracked over how much complexity he'll put up with.
USB4 V2 continues to evolve TB and further builds out protocol.
To the extent that TB features require specifically rated cables, there's a minor headache of knowing what's needed. The practical limits are due to extremely well understood physics of which there's no danger of being unlearned to the degree that you expect the chips to continue to get made. Wire heating is a direct function of current can be mediated with higher voltages. If they're doing 240W at 5V that's 50 amps! Which seems like a lot. But cable resistance can't be more that a tiny fraction of an ohm, and the connector already needs to have fantastic integrity for 50 gigahertz signaling, which is the real mystery.
The only thing that feels bad to me about this tech is the branding is dumb. It could use a bit of Steve touch.
What it looks like is that with USB4 V2 we stand at threshold of a general-purpose multicomputer interconnect technology at a bargain-basement consumer application price-point. If there's a remote DMA capability lurking in the specs then doors might open to a very interesting future, per novel arrangements of chiplets.
We again stand on shoulders of giants. To me it feels like a total heyday!
The attending morbidity on this point is odd. As if there's a lost ability to dream about new ways to personalize the tech. Maybe such dreams have dissipated because the physics are so close to magic. OTOH, we've got DALL-E and Stable Diffusion to dream for us. Whee!
Let's see I just plug my tonymacx86 bat cave post generator dongle into this usb port here, and... Doh! I didn't map that port as type 65535, garrr!
- Status