Friday, October 20th 2023

Intel 14th Gen Core Desktop Processors Don't Support Thunderbolt 5

Intel earlier this week updated its desktop processor product stack with the introduction of the 14th Gen Core "Raptor Lake Refresh" processors. During our coverage of the processor, we erroneously mentioned that the processor supports the upcoming Thunderbolt 5 connectivity standard. Intel reached out to us for a correction, which revealed an interesting detail. Apparently, 14th Gen Core desktop processors do not support Thunderbolt 5, even though Intel has their own TB5 controller design codenamed "Barlow Ridge". However, these won't arrive before Q1 2024, the currently shipping "Maple Ridge" controller has only support for Thunderbolt 4. But even once Barlow Ridge is available, they will not be compatible with "Raptor Lake Refresh" socketed desktop processors.

Intel announced the Thunderbolt 5 standard in September, around the same time, it provided a technical reveal of its upcoming Core "Meteor Lake" mobile processors. The new standard provides a generational doubling in bandwidth to 80 Gbps (per direction), but relies on an exotic new feature called Bandwidth Boost. With this feature enabled, users get 120 Gbps of bandwidth in a particular direction, and 40 Gbps on its opposite direction.
Bandwidth Boost should come in handy when, for example, you are importing a large data set from a Thunderbolt 5-based external storage device to your local machine. In such a case, the receiving (Rx or inbound) bandwidth is boosted to 120 Gbps, while reducing the transmission (Tx or outbound) bandwidth to 40 Gbps. When exporting a large amount of data to this external Thunderbolt 5 storage device, the roles are reversed, where the Tx bandwidth is boosted at the expense of the Rx bandwidth.

It's important to take a moment to understand how Intel arrives at 80 Gbps as the default per-direction bandwidth of Thunderbolt 5, even when the underlying PCIe bandwidth of PCIe Gen 4 x4 adds up to 64 Gbps per direction—Thunderbolt is a complex bus, which combines a number of links besides PCIe, such as DisplayPort, which is how it arrives at its advertised bandwidth. So 80 Gbps is really just 64 Gbps of underlying PCIe bandwidth, and 120 Gbps is really just 96 Gbps, and the current Thunderbolt 4 that boasts of 40 Gbps/direction is really just 32 Gbps/direction PCIe.

Thunderbolt and USB4 are PCIe-based serial interconnect standards. Each PCIe lane has physical wiring for Rx and Tx. The way we understand Thunderbolt 5 Bandwidth Boost to work, for the ability to push 120 Gbps through a PCI-Express 4.0 x4 connection between the "Barlow Ridge" controller and the processor's PCIe root complex, the root complex should be able to re-task both sub-lanes of two of the four PCIe lanes for either purely-Tx or purely-Rx. This probably requires some awareness at the end of the PCIe root complex that's part of the processor's uncore.
In the Intel slide above, the company details how Bandwidth Boost works at the Thunderbolt 5 link layer. Two 40 Gbps links per direction normally make up the 80 Gbps per-direction bandwidth of Thunderbolt 5. With Bandwidth Boost enabled, three out of four links are loaded toward one direction to make 120 Gbps, while the last link is in the opposite direction. As for which direction gets the Bandwidth Boost depends entirely on the application, as detailed in the examples above. You get get a 120 Gbps high-speed transmit with a 40 Gbps receive, or a 120 Gbps high-speed receive with a 40 Gbps transmit. The second image drawn by us, illustrates how the PCIe backend works between the Thunderbolt 5 host controller and the PCIe root complex.

Without this uncore-level awareness, for Thunderbolt 5 to be able to push 120 Gbps to a downstream device, there needs to be a PCIe Gen 4 x8 connection between the PCIe root and the "Barlow Ridge" host controller that can provide 160 Gbps of per-direction bandwidth, which is both undesirable and unlikely, as all prior generation discrete Thunderbolt controllers by Intel have used a PCIe x4 connection to the platform.

Intel also stated that when "Barlow Ridge" does come out in Q1 2024, it will have compatible processors at launch, without stating what those processors are. It's possible that Intel is updating its mobile processor stack, and those processors will support Thunderbolt 5 the way Intel intends it (including the Bandwidth Boost feature). Intel already provided a technical deep-dive of its upcoming Core "Meteor Lake" mobile processor, without mentioning product launch dates. The 2024 International CES (January) provides Intel with the perfect opportunity to launch new processors. Intel put out this statement as an October 19 editor's note in its newsroom article announcing the 14th Gen Core desktop processors:
While some processors in the Intel Core 14th Gen processor family will include support for Thunderbolt 5, Intel Core 14th Gen desktop processors, specifically, will not support it. The Intel Core 14th Gen launch announcement incorrectly said, "Intel Core 14th Gen desktop processors include support for … upcoming Thunderbolt 5 wired connectivity - supporting up to 80 Gbps of bi-directional bandwidth." Intel will share additional details on Intel Core 14th Gen CPUs that support Thunderbolt 5 at a later date.
The above Intel statement says that some 14th Gen processors will support Thunderbolt 5, but not the 14th Gen Core desktop processors. Intel's "Meteor Lake" mobile processor does not come with 14th Gen Core branding, but debuts a new brand nomenclature, which leaves us with only one possibility, non-socketed 14th Gen Core processors, namely "Raptor Lake Refresh" mobile (-H and -HX). With "Meteor Lake" on the horizon, Intel may push "Raptor Lake Refresh" to the -H and -HX segments much in the same way as those segments in the 10th Gen were addressed by "Comet Lake," as "Ice Lake" handled the -U and -P segments. It's very likely that these "Raptor Lake Refresh-H/HX" processors are the ones that support Thunderbolt 5 from the 14th Gen using "Barlow Ridge" controllers, besides the upcoming "Meteor Lake" processors that have "Thunderbolt 5" integrated as part of its SoC tile.

Users on the 14th Gen Core socketed desktop platform can enjoy Thunderbolt 4 using motherboards or add-on cards that have Intel's "Maple Ridge" controller that's been in the market since 2020, which provides 40 Gbps per direction bandwidth. Thunderbolt 5 and AI Boost are two features that will only make it to socketed desktop platforms with Intel's next-generation processors. We know that "Meteor Lake" is confirmed not coming to socketed desktop platforms, which leaves us with "Arrow Lake-S," bound for the second half of 2024.
Add your own comment

37 Comments on Intel 14th Gen Core Desktop Processors Don't Support Thunderbolt 5

#1
Assimilator
Intel rushing "14th gen" so hard that they're dropping worthwhile features left right and centre. Plus we're going to be in a nightmare scenario whereby some "14th gen" processors support TB5 and some don't, so consumers will be horribly confused. All because Intel has to launch SOMETHING this year, even if that something is utter rubbish.
Posted on Reply
#2
chrcoluk
Thought thunderbolt was dead, what uses it?
Posted on Reply
#3
TheLostSwede
News Editor
chrcolukThought thunderbolt was dead, what uses it?
Apple?
Posted on Reply
#4
W1zzard

seems to be pretty stable
Posted on Reply
#5
FoulOnWhite
it's a smaller display connector on a laptop
Posted on Reply
#6
TheLostSwede
News Editor
W1zzard
seems to be pretty stable
Posted on Reply
#7
Chaitanya
TheLostSwede
Thunderbolt docks are stupidly overpriced and among the photographers(around 14 are Mac users) hardly 1 of them has spent any money on a thunderbolt dock. Rest of them rely on traditional USB hubs to get USB-A ports.
Posted on Reply
#8
dyonoctis
AssimilatorIntel rushing "14th gen" so hard that they're dropping worthwhile features left right and centre. Plus we're going to be in a nightmare scenario whereby some "14th gen" processors support TB5 and some don't, so consumers will be horribly confused. All because Intel has to launch SOMETHING this year, even if that something is utter rubbish.
Not really saying that it's a good way of going about things, but I'm quite sure that the people who really have a need for TB5 will manage. Why you ask? Well even if Raptor Lake does support TB4, you can't just assume that the laptop will be equipped with it. You have to check the spec list to see if it does, and the thunderbolt version is listed in the specs list. There's a few 1500$ laptops who don't have TB at all, or only a specific SKU within the SKU got TB. Shopping for a PC has always been a clusterfuck, so if you are looking for a specific spec you have no choice but to check what you are actually getting.

And it's not like someone will make a TB5 desktop card when there isn't a single desktop CPU that support it atm :D
Posted on Reply
#9
zlobby
Well, at least they don't suffer from the lack of vulnerabilities, so I guess it balances out?
Posted on Reply
#10
bug
Processors released in October don't support standard finalized in September. Really, that's a surprise?
Posted on Reply
#11
john_
chrcolukThought thunderbolt was dead, what uses it?
In many reviews the lack of TB support is described as a (big) disadvantage for AMD based laptops. Even if it's real significance is limited, it is used in marketing.
Posted on Reply
#12
AnarchoPrimitiv
I don't understand why they had to make TB5 so complicated? It should just be a PCIe 5.0x4 link...the marketing demands it (PCIe 5.0 /TB5) and there should only be that standard (so we dont see things like we did with TB3 and dell selling laptops with a TB3 controller that only as 3.0x2 lanes) ...meaning it can't be called TB5 without 128Gbps bidirectional bandwidth to avoid the complicated nightmare that Displayport 2.0/2.1 one is (and USB 3.2) with three different tiers of bandwidth and different cables with different capabilities.

A 5.0x4 link would have brought most eGPUs to desktop parity, would allow all sorts of external monitors and high refresh, would have been plenty for consumer external storage for years, etc....
Posted on Reply
#13
TheLostSwede
News Editor
AnarchoPrimitivI don't understand why they had to make TB5 so complicated? It should just be a PCIe 5.0x4 link...the marketing demands it (PCIe 5.0 /TB5) and there should only be that standard (so we dont see things like we did with TB3 and dell selling laptops with a TB3 controller that only as 3.0x2 lanes) ...meaning it can't be called TB5 without 128Gbps bidirectional bandwidth to avoid the complicated nightmare that Displayport 2.0/2.1 one is (and USB 3.2) with three different tiers of bandwidth and different cables with different capabilities.

A 5.0x4 link would have brought most eGPUs to desktop parity, would allow all sorts of external monitors and high refresh, would have been plenty for consumer external storage for years, etc....
Because they couldn't be worse than USB4 2.0 or whatever it's called?
Also, it's something of a cable speed limit based on the amount of available data pairs before a new connector will be needed again.

Also, the asynchronous transfer mode will most likely just be for DP signals anyhow.
Keep in mind that Thunderbolt maxes out at USB 3.2 Gen 2 10 Gbps speeds, yet Intel forced a 20 Gbps mode for the USB4 spec that no-one really wanted.
Posted on Reply
#14
bonehead123
AssimilatorAll because Intel has to launch SOMETHING this year, even if that something is utter rubbish
^^THIS^^
I think intel should just change their name to "Clusterf*ck.com" and be done with it :D
Posted on Reply
#15
Dr. Dro
Of course it doesn't, Intel should have been honest and told everyone the truth: their 14th gen CPUs aren't new and aren't worthy of being called a generation, they're anything but that! Without as much as a new stepping, let alone any redesign or change, anyone who's buying a 14900K today could have done so as far back as January when Intel released the 13900KS, and could have a chip that comes at 95% of the clock speed a year ago. They're all the same.
Posted on Reply
#16
Tek-Check
AnarchoPrimitivI don't understand why they had to make TB5 so complicated? It should just be a PCIe 5.0x4 link...
It cannot be PCIe Gen5 x4, as this link at 128 Gbps, 32 Gbps per lane, exceeds physical capability of each TX/RX data lane in TB5 PHY, which is 20 Gbps. It needs to be PCIe Gen4 x4.
TheLostSwedeAlso, it's something of a cable speed limit based on the amount of available data pairs before a new connector will be needed again.
This.
Posted on Reply
#17
Nanochip
Why not just use a pcie5.0x4 root complex? The usb4v2 spec has a revision to specifically allow for pcie5.0. Seems like that would’ve been easier and cleaner to allow for up to 128 Gbps bandwidth.

from section 11.2 (Version 2.0 with Errata and ECN through June 29, 2023)
“An Internal PCIe Port shall be compatible with the PCI Express® Base Specification, Revision 4. An Internal PCIe Port may be compatible with the PCI Express® Base Specification, Revision 5.0.
Posted on Reply
#18
Assimilator
bugProcessors released in October don't support standard finalized in September. Really, that's a surprise?
Processors that don't need to be released in October because they offer no advantage over their "predecessors" in any way, nor are their any new processors going to be released by their competitor, so it makes zero sense to release them now when you could release then in January and have them include a standard finalized in September that makes them a somewhat compelling buy.

There, fixed.
Posted on Reply
#19
Nanochip
AssimilatorProcessors that don't need to be released in October because they offer no advantage over their "predecessors" in any way, nor are their any new processors going to be released by their competitor, so it makes zero sense to release them now when you could release then in January and have them include a standard finalized in September that makes them a somewhat compelling buy.

There, fixed.
Intel has been releasing desktop CPUs in the fall for a few years now. Alder lake. Raptor lake. Raptor lake.1 aka 14th gen.
Posted on Reply
#20
Super Firm Tofu
Dr. Droanyone who's buying a 14900K today could have done so as far back as January when Intel released the 13900KS, and could have a chip that comes at 95% of the clock speed a year ago. They're all the same.
Except now they can do it for $200-$300 less.

I guess you lose the "S" which obviously is the only way to maintain your l337 status and be xtreem.

It's horrible that intel has offered (minimally)more for less money. Dirty bastards.
Posted on Reply
#21
pressing on
Super Firm TofuExcept now they can do it for $200-$300 less.

I guess you lose the "S" which obviously is the only way to maintain your l337 status and be xtreem.

It's horrible that intel has offered (minimally)more for less money. Dirty bastards.
At least Intel are now done with generations. In future their chips will just have numbers like 135, 155 and 185. In a couple of years will people be arguing that the 205 isn't really a 200 series it's just a refreshed 185...
Posted on Reply
#22
ir_cow
Thunderbolt 5 aka Display Port 2.0
Posted on Reply
#23
trsttte
ChaitanyaThunderbolt docks are stupidly overpriced and among the photographers(around 14 are Mac users) hardly 1 of them has spent any money on a thunderbolt dock. Rest of them rely on traditional USB hubs to get USB-A ports.
And that's the real tragedy, until thunderbolt accessories become affordable so anyone feels they're worth buying they'll continue to be a niche in the market, which will make them continue to be expensive. About the only real market for thunderbolt are offices and professional environments where price doesn't really matter (it's a write off anyway) but the feature set does.
AnarchoPrimitivthere should only be that standard (so we dont see things like we did with TB3 and dell selling laptops with a TB3 controller that only as 3.0x2 lanes)
That was already fixed with thunderbolt 4
AnarchoPrimitivA 5.0x4 link would have brought most eGPUs to desktop parity, would allow all sorts of external monitors and high refresh, would have been plenty for consumer external storage for years, etc....
Only a problem if you use the eGPU with the internal monitor, otherwise it's fine.
Dr. DroOf course it doesn't, Intel should have been honest and told everyone the truth: their 14th gen CPUs aren't new and aren't worthy of being called a generation, they're anything but that! Without as much as a new stepping, let alone any redesign or change
There's at least one change - the name - and that's the only one that really matters. OEMs demand a new processor so they can refresh their designs and claim all improvements imaginable no matter how stale the performance has been for the last couple years.
Posted on Reply
#24
bug
AssimilatorProcessors that don't need to be released in October because they offer no advantage over their "predecessors" in any way, nor are their any new processors going to be released by their competitor, so it makes zero sense to release them now when you could release then in January and have them include a standard finalized in September that makes them a somewhat compelling buy.

There, fixed.
This was always going to be a refresh. If you start messing with the silicon, then it's not a refresh anymore.

And they certainly offer advantages: reduced power draw, minor speed bumps and 4 more E-core for the 14700k. Quite hefty for a refresh, imho, even if far from what one would expect from a new generation (which these are not).
Posted on Reply
#25
Broken Processor
Why? I use TB it's versatile and handy for display if I was going to buy I'd be looking for it.
Seems on par for the 14th gen launch, this launch shows what's wrong with the industry.
Would have been safer calling it 13+ or something but it absolutely didn't deserve to be called a new gen.
Posted on Reply
Add your own comment
Nov 21st, 2024 12:56 EST change timezone

New Forum Posts

Popular Reviews

Controversial News Posts