Intel executive accidentally reveals that Thunderbolt 5 will achieve 80 Gbps thanks to...

nanoguy

Posts: 1,355   +27
Staff member
Something to look forward to: Intel executives may be good at overseeing technological innovations, but sometimes they share their enthusiasm on social media and unintentionally reveal bits about ongoing projects that aren't supposed to be public yet. Such is the case of Gregory M. Bryant, GM of Intel's Client Computing Group, who inadvertently revealed on Twitter what the company has in store for the next generation of Thunderbolt.

Thunderbolt 4 was made official last year with the debut of Intel's 11th-gen Core Tiger Lake CPUs, bringing a number of improvements in terms of security and display support, while keeping the same 40 Gbps bandwidth as Thunderbolt 3. However, a recent leak indicates it won't be long before Thunderbolt 5 comes out.

During a recent visit at Intel's research and development labs in Israel, Bryant took a series of photos and shared them on Twitter. In a now deleted tweet, Bryant posted a photo of what appears to be a Thunderbolt 5 demo setup and a wall poster detailing the technology behind it.

The poster read "80G PHY Technology," which suggests the new standard will offer a theoretical bandwidth of up to 80 Gbps -- double that of Thunderbolt 4. There's also a line that reads "USB 80G is targeted to support the existing USB-C ecosystem," indicating that Thunderbolt 5 will continue to use the same USB Type-C connector as its predecessors.

The magic behind the improved transfer speeds in Thunderbolt 5 are apparently made possible thanks to a "novel PAM-3 [Pulse Amplitude Modulation 3-level] modulation technology." As explained by AnandTech's Ian Cutress, this refers to a data transmission protocol where the data line can carry a 3-bit data signal encoded with logical +1, 0, and -1 values. For instance, "000" is "-1" followed by "-1", while "001" is "-1" followed by "0," and so on.

At 1.5 bits per symbol, PAM-3 sits right between other encoding methods such as NRZ and PAM-4 that achieve 1 and 2 bits per symbol, respectively. But more importantly, this allows PAM-3 to achieve a higher bandwidth that of NRZ without the limitations of PAM-4, which requires more complex equipment due to signal quality issues. In other words, we could see cheaper cables for the next Thunderbolt standard.

Intel is supposedly testing a 6nm chipset for Thunderbolt 5, which means the company will likely tap TSMC to make it. At this point, all we have is this unintentional sneak peek into the new technology, but Intel seems to be onto something here, especially as it retains compatibility with the growing USB-Type C ecosystem.

Permalink to story.

 
I'd care more about Thunderbolt if there were PCIe cards available to add it to an existing system.
I just did a quick Google search and found multiple pcie Thunderbolt cards.

I'm just glad they are keeping the usb c connector because I do not relish the thought of having to buy even more usb cords I'll inevitably lose.
 
About time. 40Gb was already limiting factor for high data transfers.

Now all what Intel has to do is to finally open this standard for the masses... and AMD. Apple is gone from their list so they cannot just lock this behind royalties because nobody will give a 2fs about TB5 then.


 
And how many -really- care about Thunderbolt?

My RTX 3080 have PAM4 signalling :joy:

Can't wait to see Alder Lake performance tho...
Most people probably aren't even aware of the benefits of Thunderbolt, that's why they don't care. I, for one, don't understand why any sane person on Earth would prefer more cables over less.
 
Most people probably aren't even aware of the benefits of Thunderbolt, that's why they don't care. I, for one, don't understand why any sane person on Earth would prefer more cables over less.
I’m all usb C and I still have the same amount of cables because I like to charge things at the same time. It’s a hard life..
 
Most people probably aren't even aware of the benefits of Thunderbolt, that's why they don't care. I, for one, don't understand why any sane person on Earth would prefer more cables over less.
The handshake issues are enough to keep me away. I don't need it at all - USB-C does fine for my needs.

We had tons of issues with Thunderbolt docks (or more like port replicators) in my company. To the point where we replaced almost all with USB-C ones, not a single issues since.
 
I just did a quick Google search and found multiple pcie Thunderbolt cards.

I'm just glad they are keeping the usb c connector because I do not relish the thought of having to buy even more usb cords I'll inevitably lose.
A quick google search shows those thunderbolt cards require AN INTERNAL HEADER, which means they are NOT adding thunderbolt to a system, they are providing an interface for a system taht is already on the motherboard.

To date there are 0 cards that allow you to add thunderbolt to a system that does not already have it, say a ryzen system or a coffee lake/haswell system. I'd love for you to link oen to prove me wrong, but google it seems says they do not exist.
 
About time. 40Gb was already limiting factor for high data transfers.

Now all what Intel has to do is to finally open this standard for the masses... and AMD. Apple is gone from their list so they cannot just lock this behind royalties because nobody will give a 2fs about TB5 then.

Google "USB 4"

You're welcome.
 
I see a pattern here. It took me several weeks on my quantum computer but I now can predict with great accuracy that Thunderbolt 6 will ... drum roll please ... hit 160Gb/s.

No please hold the applause, seriously just donate to a charity of you liking.
 
The handshake issues are enough to keep me away. I don't need it at all - USB-C does fine for my needs.

We had tons of issues with Thunderbolt docks (or more like port replicators) in my company. To the point where we replaced almost all with USB-C ones, not a single issues since.

What handshake issues - please elaborate as I am looking to buy TB laptops.
 
Back