Comments Locked

76 Comments

Back to Article

  • ikjadoon - Wednesday, July 8, 2020 - link

    The type-C port identification debacle continues full-steam ahead, muddling any attempt for docks to proliferate (and lower in cost). The USB-IF Board (Apple, Intel, HP, Microsoft, Renesas, STMicroelectronics, and Texas Instruments) has done next to nothing useful for the better part of a decade, i.e., USB-C was finalized in 2014.

    At this point, I'd instead appreciate a OS-based menu that detailed each type-C port's capabilities in a silly matrix chart. Someone mentioned this and it seems...almost too logical.

    It's over for consumer understanding of type-C: it's utterly unsalvageable. Maybe type-D in 2030. But for people who just want compatible docks / parts, operating systems should take over with a simple menu accessed in their Settings app.

    That would also solve the # / bitrate display issue (the OS knows which GPU), there are no godforsaken logos every OEM seems to hate, and you can't lose it like a product info webpage 404'ing years later or a lost manual.

    And for TB3, it could detail whether x2 and some are x4, whether dedicated silicon or partially integrated (Ice Lake) or fully integrated (Tiger Lake) [lower latency w/ more integration].

    Your OEM doesn't care.
    USB-IF doesn't care.
    Dock manufacturers are screwed every which way.
    Consumers are priced out of the market.

    Just another silly and greedy reason why work from home should've been an easier transition, but just sheer greed has prevented even a little bit of industry collaboration.

    cc: Microsoft, Apple, Google
  • Spunjji - Wednesday, July 8, 2020 - link

    +1 to this. I had such high hopes for USB-C, but it was all downhill once Thunderbolt piggybacked on the connector.

    I'm not blaming Intel for this specifically - the USB-IF have done their best to make USB 3's various sub-standards into a complete joke - but that splitting point marked the beginning of the end.
  • repoman27 - Wednesday, July 8, 2020 - link

    The only reason the USB-IF has the Type-C connector at all is because Apple wanted a small reversible connector like Lightning that they could use for Thunderbolt, DisplayPort, USB and USB Power Delivery.
  • Strunf - Thursday, July 9, 2020 - link

    I doubt that, Apple is known to go their own ways just to keep overcharging for their certified connectors, they even went after all those that were using their proprietary connectors in non certified accessories, also the iPhone still doesn't use USB-C.
    The USB-IF made the USB-C probably cause of market pressure, Samsung and others for sure were not happy seeing how Lightning was so much better.
  • rahvin - Friday, July 10, 2020 - link

    USB-C was only partially apples fault but mostly for designing the widely adopted and horrible design standard that was micro-usb, a connector designed to be so weak the connector would fail under routine use after a couple hundred inserts. They were trying to find a better design than micro-usb which was so strong it broke the board rather than the cable and failed miserably the other direction.

    USB C was developed to solve the quantum USB-A insert problem where it always takes 3 attempts to insert it the right direction. It also dealt with the USB-IF desire to increase the wire count to accommodate future needs by raising the number of pins and wires to 24 (which allowed the inclusion of alt-mode uses like video delivery, power delivery and all the other additions.

    No the problem with USB-C has always been that USB-IF allowed the cable manufacturers to deliver cables and devices that only partially met the spec. The result is you need a doctorate in USB-C to understand what part of the standard each cable and device using USB-C supports. Today as a result of USB-IF bowing to marketers you have a standard with cables that don't support anything higher than USB-2 because 2/3rds of the wires are missing.

    The result of all this is the table above of all the different icons needed to indicate what a USB-C device is capable of, and the icons aren't mandatory. So get your diploma in USB-C if you want to buy something supporting it because god knows the manufacturers lie their asses off and that is all the USB implementation forum's fault.
  • rahvin - Friday, July 10, 2020 - link

    I mistyped and didn't proof read properly,

    Micro-USB was a badly designed solution to the Mini-USB standard damaging/destroying circuit boards when the cable was violently removed. Apple went too far the other direction and designed a cable connector that was so weak it failed under routine use. A typical micro-usb cable has to be replaced every few hundred insertions. USB-C was meant to find the right balance, and solve the directionality problem that Type-A suffered while raising wire count and allowing for the alt-modes.

    Overall I think USB-C is a great connector, the problem was not forcing compliance with the whole specification to use the connector.
  • repoman27 - Wednesday, July 8, 2020 - link

    Since 2014, the USB-IF has, among other things, released revisions 1.1, 1.2, 1.3, 1.4, and 2.0 of the USB Type-C Cable and Connector Specification, revision 2.0 versions 1.1, 1.2, 1.3 and revision 3.0 versions 1.0, 1.0a, 1.1, 1.2, and 2.0 of the USB Power Delivery Specification, revision 1.0 of the USB 3.2 Specification, and version 1.0 of the USB4 Specification. So not exactly nothing. And given that it generally takes several years to go from spec to silicon, a lot of what they have done is not yet visible to consumers.

    There are obviously cases where vendors have failed to properly implement the USB Type-C and USB Power Delivery specifications leading to... suboptimal user experiences.

    However, I think you may be looking at "consumer understanding" differently than the USB-IF. Their goal is more along the lines of "if it fits, it works". 99% of the time, all the end user wants is data and/or power to flow between two connected devices. Whether you're transferring data or charging at the maximum possible rate is rarely the primary concern. In which case, any passive USB Type-C to Type-C or Type-C to legacy cable that actually follows the spec will just work if the connectors on both ends fit.

    If a device manufacturer doesn't advertise a product as supporting USB Power Delivery or specific Alternate Modes, you should probably assume that USB port is just a USB port. And like any USB port, that means you should also assume it only supports USB 2.0 and not more than 500 mA (2.5 W) over the VBUS. Anything beyond that is a bonus. The number of workflows that actually depend on SuperSpeed USB transfer rates in excess of 435 MB/s (the practical limit of USB3 Gen 1 x 1) is essentially zero. Even when it comes to eGPU usage or digital video workflows, Thunderbolt 3 PCIe throughput of more than 1400 MB/s (the limit when the Thunderbolt controller is connected to the host / device via a PCIe Gen3 x2 or Gen2 x4 link) is rarely a necessity.

    Active cables, on the other hand, are an actual problem. People really don't seem to understand why certain cables only support certain signaling modes, and vendors have done very little to make this clear.
  • Maxed Out - Wednesday, July 8, 2020 - link

    While you are absolutely correct, there is no stopping stupid.

    We had color-coded and keyed connectors and people couldn't plug in things properly.

    This is coming from a former onsite computer technician, who has built and supported everything from HTPC to Data Centers.
  • rhysiam - Wednesday, July 8, 2020 - link

    There will always be a portion of the user base who need everything done for them (a certain tech-support call logging the PC's broken "coffee holder" (=CD drive) comes to mind), but there are many competent users for whom @ikjadoon's post is bang on IMHO. Let's say in a few years my mate wants a dock for his laptop which provides two external displays, power deliver, 10GBe and a couple of USB4 connectors: Which dock should he buy? He shouldn't have to ask me and I shouldn't have to spend significant time going digging up product manuals, confirming compatibility and checking he's got the right cables for that all to work.

    I agree a small portion of the user base wouldn't have got it whatever was done, but some basic colour differences, or symbols, or something - mandated by the standard - would have a gone a long way here!
  • BedfordTim - Thursday, July 9, 2020 - link

    Exactly. When you go to buy something the exact specification is incredibly hard to find. Even basics like does it support PD are tucked away in a manual you have to register to view.
  • jospoortvliet - Thursday, July 9, 2020 - link

    Yeah, it has to be checked on the side of the laptop, the cable AND the dock. It is absolutely insane - a USB 4 logo should mean a baseline of functionality - display, power etc - not this mess it is now.
  • trenzterra - Tuesday, July 14, 2020 - link

    Even manufacturers get it wrong sometimes. I bought a HP Pavilion laptop on the understanding the spec page and manual clearly showed support for PD charging. It didn't. Turned out they copy pasta'ed the specs sheet from an Envy laptop.
  • voicequal - Wednesday, July 8, 2020 - link

    TB4 seems to be going out of its way to be backward compatible with all other uses for USB-C. I'm not sure what your post is advocating on the HW side? Investing in one connector makes a lot of sense. See how far Ethernet has gone with 25 years of RJ-45: 10M/100M/1G/2.5G/5G/10G + PoE. Yes, better OS support and capabilities flags for multimodal interfaces can always help.

    Dock vendors are incentivised to highlight what they can do and be silent on what they can't, leaving the consumer to decipher the gaps. TB has always been a premium product, so you'll most likely to get the fullest capabilities from a TB dock, but rightly at higher cost than barebones USB hubs.
  • saratoga4 - Wednesday, July 8, 2020 - link

    The thunderbolt port logo looks way too much like a charging/power supply icon for something that is suppt to be a data cable.
  • Tomatotech - Wednesday, July 8, 2020 - link

    FYI, it's also the charging port. (but only if you get the correct port ... )
  • Tomatotech - Wednesday, July 8, 2020 - link

    Edit: On Mac laptops with TB3, you can charge on any port. On Windows... flip a coin, especially if you're wondering which port to use on a dead laptop. (the jokes just write themselves ...)
  • jeremyshaw - Wednesday, July 8, 2020 - link

    Outside of TB3, some laptops with USB-C charging won't charge if the battery is completely flat. Pathetic. ASUS G14, if you are curious - only the barrel jack charger will work if the battery is below 3% or so.
  • Mitch89 - Wednesday, July 8, 2020 - link

    On Windows laptops there’s no guarantee any of the USB-C ports charge the machine, nor support DisplayPort Alt.

    My wife’s ASUS has a USB-C, but it’s no more useful than the USB-A port.
  • xenol - Wednesday, July 8, 2020 - link

    It also supplies power. At this point it's ingrained in people's minds that USB is also a power supply port for their devices.

    The charging port is an issue, but it seems most laptops still use a typical barrel connector which I'd prefer.
  • jeremyshaw - Wednesday, July 8, 2020 - link

    Up to 3GB/s (~24Gbps)? May you push them on that? It seems the annoying, seemingly arbitrary PCIe tunneling bandwidth limit is still basically the same (it was about ~2.75GB/s before, still a fraction of the total "40Gbps" they claim for TB3. It's all for naught if the only thing that can use all 40Gbps is Displayport...
  • Techtree101 - Wednesday, July 8, 2020 - link

    Total is 80 Gb/s for single direction connections like DisplayPort 2.0.
  • repoman27 - Wednesday, July 8, 2020 - link

    This isn't an arbitrary limitation. The discrete controllers only have a PCIe 3.0 x4 connection to their host or device. The signaling rate for PCIe Gen3 is 8 Gbit/s, which x 4 = 32. But then there's 128b/130b encoding which reduces the effective bandwidth to 31.508 Gbit/s. PCI Express is a packet based protocol, so in addition to payload data, there's also framing, headers, and CRC, plus packets used by the transaction and data link layers for management. On most systems, this protocol overhead brings the payload capacity down to ~26.6 Gbit/s or 3325 MB/s.

    The Thunderbolt 3 link is two bonded full-duplex channels operating at a signaling rate of 20.625 Gbit/s with 64b/66b encoding, which works out to a full 40 Gbit/s for data. However, the controller has to take the PCIe packets from the host / device, further encapsulate them and then transport them over the Thunderbolt link. The protocol overhead (which I believe may be further increased by the way channel bonding is implemented) results in peak Thunderbolt 3 PCIe payload throughput of ~2750 MB/s. However, that is probably due in part to limitations of the lane adapters rather than the Thunderbolt channel being completely stuffed.

    Tiger Lake will be the first Intel platform to support PCIe Gen4, however it also has integrated Thunderbolt 4. The Maple Ridge controllers are probably PCIe Gen3 only for the sake of keeping power requirements in line. But who knows, maybe they'll suddenly support PCIe Gen4 once Rocket Lake is released.
  • jeremyshaw - Wednesday, July 8, 2020 - link

    I do agree with the overhead of PCIe on the PCIe side of things, however, Intel disagrees with your assessment of TB3's own tunneling. Of course, in extremely un-Intel fashion, they have almost NO technical data on this publicly released, despite it being an optional part of the USB4 protocol and the basis of the DP2.0 protocol. Basically, the overhead of packaging a PCIe-only stream will not somehow reduce the B/W of TB3's 40Gbps down to ~22Gbps. Intel's own claim (shown by Anandtech on their 19FEB2019 review of an eGPU box) is 22Gbps max for PCIe, arbitrarily. This is why I'm not buying the overhead, since TB3 (in the same Intel claim in that Anandtech review) shows it can handle ~18Gbps PCIe + ~22Gbps video (specifically, their claim is 5k@60Hz with 10bit color - the data rate for that is calculated to ~27Gbps... I can only assume they really meant 8bit color, which is closer to 22.18Gbps).
  • repoman27 - Wednesday, July 8, 2020 - link

    The PCIe Gen3 x4 back end limits PCIe payload throughput to 26.6 Gbit/s. That's the max any PCIe Gen3 x4 device can muster (unless you bump up the max TLP payload size).The PCIe protocol / lane adapters in Intel's Thunderbolt 3 controllers can only manage 83% of that rate, but they're also dealing with round-trip latencies on the Thunderbolt link that are orders of magnitude higher than a typical PCIe link manager has to deal with and having to send multiple copies of certain types of control packets to accommodate Thunderbolt channel bonding. As I said, the limitation isn't because Thunderbolt bandwidth is saturated, but it's also not arbitrary. That's simply as fast as she'll go.

    The DisplayPort tunneling strips out all of the bit stuffing and most of the blanking from the regular DP stream and then encapsulates it, so you need to take that into account with your bandwidth calculations. The USB4 Specification is free to download from the USB-IF if you want to go plumbing the depths of how the protocol works.
  • jeremyshaw - Wednesday, July 8, 2020 - link

    I already downloaded the USB4 spec on the day it was released, and read through it. Unless if it has massively changed since then (it hasn't), it talks a lot about implementations and some theory, but TB3 is largely handwaved as (paraphrasing) "follow TB3 protocol," etc. For example, the DP tunneling aspect of the TB3 compatibility mode (NOT the same as DP-Alt mode)? It has plenty of sections that devolve into "Follow DP 1.2 spec; follow DP1.4a spec." Where? Oh, right behind that paywall. Speaking of the DP spec, it doesn't state anywhere (nor is it inferred from reading any of the specs) that DP is stripped and reencoded/encapsulated/etc. Same is true of the TB3 implementation. Almost all of the data you see is about implementing the surface layer of TB3 compatibility (bringup/initialization). Everything else is still (paraphrasing) "in the actual data stream, this section is TB3 compatible XXXXXX." Hey, maybe I'm wrong and the USB-IF manual is massively different, nowadays. I recently downloaded it again and read the redline version, and I didn't see any meaningful or insightful improvements in the area of TB3.

    W.r.t. blanking, timing, etc for DP, HDMI, etc I'm already well aware of it. Intel's own calculations are still wrong. 22Gbps will not get you a tunneled DP with 5k, 60Hz, 10bits. It would barely work at 8bits. If one uses raw data calculations, it is ~26.5 Gbps for 5k, 60Hz, 10bits - this is easy for anyone to verify, instead of trusting Intel's claims.

    As for the overhead, I'm still not seeing the explanation on how 26.6Gbps becomes 22Gbps out of a 40Gbps channel, especially when that 40Gbps channel is capable of saturating itself unidirectionally, with mixed PCIe and DP tunneling. Even if there is some sort of fundamental mis-timing and misalignment that just saps everything away (which again, would be easier to see if we had any sort of technical data on these, beyond tiny snippets of hints from Intel/Partners, and vague, dismissive references from other standards bodies), it would not explain away the ability to have a mixed PCIe and DP load reach 40Gbps.

    Unless if anyone has any hard technical data on TB3's internal logic and functionality w.r.t. the PCIe bandwidth limitation, we're still at an impasse. The limit is still arbitrary (unless if we have differing definitions of arbitrary) and we are still in the dark as to why TB3 has a 40Gbps link with only ~22Gbps PCIe passthrough.
  • repoman27 - Wednesday, July 8, 2020 - link

    Apologies. I agree that it is assumed that you are a VESA member and have access to the DisplayPort specs, and that everything pertaining to TBT3 compatibility in that document is vague af. I also agree that that slide from Intel marketing showing the bandwidth calculations for Thunderbolt 3 is clearly flawed, and that without lossy compression you can't make a 5120 x 2880 @ 60 Hz, 10 bpc stream less than 26.54208 Gbit/s. However, I do kinda feel like the protocols themselves are pretty clearly defined in their respective specifications.

    DisplayPort Tunneling and Alternate Mode are completely different. The former is tunneled over USB4 / TBT3, the latter is native signaling. USB4 DP tunneling is almost the same as TBT3 DP tunneling with a few differences.

    From section 10.5.1 SST Tunneling:

    "When a DisplayPort SST Main-Link is mapped onto USB4, the continuous Main-Link data stream
    is encapsulated into Tunneled Packets. Before encapsulation, all the Stuffing Symbols (within a
    TU and during the blanking periods) are discarded by the DP IN Adapter. Stuffing Symbols are
    recreated by the DP OUT Adapter when the Main-Link data stream is extracted from the
    Tunneled packets. In order to enable accurate reconstruction of the Stuffing Symbols at the DP
    OUT Adapter, the DP IN Adapter includes a Fill Count field in each Tunneled Packet. The Fill
    Count field specifies the number of Stuffing Symbols that were discarded immediately preceding
    the packet."

    And from section 13.8 DisplayPort Tunneling:

    "TBT3-Compatible mode is Non-LTTPR mode as defined in Chapter 10 with the modifications defined in this section..."

    But to bring this back to PCIe tunneling, regardless of the Thunderbolt link bandwidth, PCIe bandwidth between any single device and the host cannot exceed what the PCIe link between the controller and host / device is capable of. It's not that hard to calculate the theoretical maximum throughput of a PCIe Gen 3 x4 link and confirm those calculations via empirical data. It's also not hard to look at the real-world data for peak Thunderbolt 3 PCIe throughput and see that it tops out right around 2750 MB/s or 22 Gbit/s. Strangely enough, Intel marketing used that number in their slide and it proves out in real life. I see three options here:

    1. Certain aspects of the Thunderbolt 3 protocol limit PCIe tunneling throughput to less than what the back-end of the controller is capable of.

    2. Intel's implementation of PCIe tunneling in their discrete Thunderbolt 3 controllers is not capable of sustaining PCIe tunneling rates greater than 22 Gbit/s and Intel is not keen on discussing why.

    3. Intel is arbitrarily limiting PCIe tunneling throughput from ~26 to 22 Gbit/s.

    While we may still be in the dark, I certainly feel that option 3 is by far the least likely.
  • Eliadbu - Wednesday, July 8, 2020 - link

    After digging a bit I found an early thunderbolt 3 brief that explains bit about priorities and data rate. While the connector is definitely capable of 40 Gbps, there is an illustration that shows that the maximum data rate as If you transfer just data is 22 Gbps each direction and other bandwidth is not used. Makes scratch my had and wonder why the specification limited for data transfer while if you pass video it can do much more. I guess there are maybe some limitations deep down that limit the data rate from pcie.
  • Deicidium369 - Thursday, January 14, 2021 - link

    TB4 is integrated on the Tiger Lake CPU - not connected as an external controller. Intel NUC11 Pro has TB4 and PCIe4 for the M.2 SSD.

    https://ark.intel.com/content/www/us/en/ark/produc...
  • Silma - Wednesday, July 8, 2020 - link

    Wasn't USB 4 supposed to be fully compatible with Thunderbolt 3 ?
  • Valantar - Wednesday, July 8, 2020 - link

    It is. One of the slides in this article even states so.
  • repoman27 - Wednesday, July 8, 2020 - link

    No.

    USB4 is only conceptually based on Thunderbolt 3; it differs materially in terms of specific implementation details. Most notably, the two protocols actually use slightly different signaling rates and encoding schemes. Thunderbolt 3 interoperability is optional for USB4 hosts and devices and implemented via a Type-C Alternate Mode, same as it was under USB3. Interoperability is only required for docks and to a certain degree on hubs.

    From the specification:

    A USB4 host or USB4 peripheral device can optionally support interoperability with Thunderbolt 3 (TBT3) products.

    A USB4 hub is required to support interoperability with Thunderbolt 3 products on all of its DFP. A USB4-Based Dock is required to support interoperability with Thunderbolt 3 products on its UFP in addition to all of its DFP.

    When interoperating with a TBT3 product, Thunderbolt Alt Mode is established on the link between products. The USB Type-C Specification describes how a USB4 product negotiates and enters Thunderbolt Alt Mode.

    Also from the spec:

    Thunderbolt™ is a trademark of Intel Corporation. You may only use the Thunderbolt™ trademark or logo in conjunction with products designed to this specification that complete proper certification and executing a Thunderbolt™ trademark license.

    Intel is still gonna Intel.
  • serendip - Thursday, July 9, 2020 - link

    So TBT3 is a superset over USB, just like DisplayPort Alt Mode? The device is supposed to follow USB4 for full compatibility and add alternate modes if needed.
  • Deicidium369 - Thursday, January 14, 2021 - link

    TB4 spec contains both TB3 and USB4.

    you can have USB4 without TB3/4
  • Spunjji - Wednesday, July 8, 2020 - link

    Is it just me or is Thunderbolt 4 looking increasingly like a rehash of Thunderbolt 3, just guaranteed to work at full speed and with some deeper vendor lock-ins?

    The "one port to rule them all" concept could be beautiful, but as it is, it's actually just adding to the confusion because it's impossible to differentiate from Thunderbolt 3 and will probably not be ubiquitous.
  • DigitalFreak - Wednesday, July 8, 2020 - link

    It's TB3 with a 1m longer cable and optional stuff now required. At least USB increases it's speed with every version.
  • chaos215bar2 - Wednesday, July 8, 2020 - link

    There are plenty of 2m TB3 cables which support 40Gb/s data and 100W charging. What's different, if I'm reading correctly, is that Thunderbolt 4 cables must support full backwards compatibility. That could just be a change to the active hardware used on 2m cables.

    What's less clear is whether the longer optical cables will be fully compatible. Based on what's said here, I'd expect a 50m Thunderbolt 4 cable to also work as a really expensive USB 2.0 cable.
  • DigitalFreak - Thursday, July 9, 2020 - link

    I just saw that there are 2m TB3 active cables now that can do the full 40Gbps. The last time I looked into it you could still only do a max of .5 (or 1?) meters @ 40Gbps. Those active cables are expensive af though. Don't even get me started on the fiber ones.
  • repoman27 - Wednesday, July 8, 2020 - link

    Thunderbolt 4 is USB4 with full support for the Thunderbolt 3 and DisplayPort Alternate Modes. This is significant in that Thunderbolt 3 controllers do not support USB4 at all. The two protocols are not the same.

    The differences between USB4 and TBT3 are discussed starting at page 68 in this slide deck: https://www.usb.org/sites/default/files/D2T1-3%20-...
  • Deicidium369 - Thursday, January 14, 2021 - link

    Well that would be your take...Through your AMD Colored Glasses

    TB4 is a certification. it certifies that the implementation of TB3 and USB4 meet a certain level of functionality. No vendor lock in - Nothing stopping AMD board vendors from adding TB4 support - nothing.

    A single cable - a TB4 cable - will operate at full speeds for both TB3 and USB4 over the C connector (40Gb/s)

    Apple and Intel support TB - so already ubiquitous - and only a small niche player lacks it
  • PaulHoule - Wednesday, July 8, 2020 - link

    Whoever made that "How Thunderbolt 4 is different from other solutions?" slide should make luxury car commercials.

    My concern is this: I have used Mac and Dell computers with Thunderbolt support and I'd say the experience is good on the mac and worse on Windows. A newer Dell I use came with a Thunderbolt dock that "just works" (charging, display, network and HID) but I have had a hard time finding pairs of machines and docks that work completely. Often everything works except the network. My Alienware (also Dell) did OK with a Thunderbolt monitor I tried, although it took much longer to configure itself than macs did.

    So now I have a "four plug" docking solution at home which works with every laptop I have tried, which involved trying quite a few different devices, giving up on Logitech/Plantronics dongles for Bluetooth, ... It supplies 250W, but it does not have the one-plug simplicity that i wish for.
  • schizoide - Wednesday, July 8, 2020 - link

    Originally TB4 was rumored to be 4 lanes of PCIe 4.0, which would double the bandwidth. That extra bandwidth would have been very useful for eGPU purposes, particularly if you want to use your laptop's integrated display. Pity it turned out to be such a minor update.
  • Techtree101 - Wednesday, July 8, 2020 - link

    Correction: It will offer up to 80 Gb/s when in a single direction. 40 GB/s is for bi-directional use. So, for example, DisplayPort 2.0 will be using it at 80 Gb/s.
  • nandnandnand - Wednesday, July 8, 2020 - link

    Given that, it should be easy to figure out "what refresh/bit-rate/chroma" it will support. Unless something has gone horribly wrong and TB4 doesn't support these:

    https://en.wikipedia.org/wiki/DisplayPort#DP_2.0_c...
  • repoman27 - Wednesday, July 8, 2020 - link

    Thunderbolt 4 controllers don't support DP 2.0 because no existing GPUs support DP 2.0. I believe they will all be DP 1.4a HBR3 / DSC 1.2a capable though.

    @nandnandnand The refresh/bit-rate/chroma is dependent firstly on the capabilities of the GPU and connected display. Until such time as a DP 2.0 capable GPU and scaler / TCON exist, anything beyond HBR3 with DSC won't be achievable anyway.
  • Techtree101 - Wednesday, July 8, 2020 - link

    Read: https://appleinsider.com/articles/19/06/26/new-dis...

    Yes, it uses TB3, but this will apply directly to TB4.
  • repoman27 - Wednesday, July 8, 2020 - link

    I am pretty well acquainted with the actual specifications. DisplayPort 2.0 leveraging the USB4 / Thunderbolt 3 PHY does not magically make the necessary DisplayPort 2.0 IP appear in the display controller blocks of GPUs, displays, or discrete Thunderbolt controllers. If there is no way to get a DP 2.0 UHBR signal into or out of a Thunderbolt controller (and on to the host or device), then the capabilities of the Thunderbolt PHY aren't really relevant. The maximum end-to-end DP link you can establish is still limited to HBR3.
  • Techtree101 - Wednesday, July 8, 2020 - link

    Are you saying then that future devices will not be able to use DP 2.0 and Thunderbolt 4 together, at least not at the DP 2.0 max bandwidth, ever?
  • repoman27 - Wednesday, July 8, 2020 - link

    No. What I'm getting at is that although Intel may someday introduce Thunderbolt 4 products that do support DP 2.0, it is highly unlikely that the Maple or Goshen Ridge controllers can or ever will.
  • Techtree101 - Wednesday, July 8, 2020 - link

    That's what I understand as well. My apologies if there was confusion here. My point in responding was that with the proper future hardware, DP 2.0 and TB4 will work together (at some point) to gain you that bandwidth.

    The article says: "Thunderbolt 4 will offer speeds up to 40 Gb/s", which then goes on to talk about existing upcoming controllers, a separate topic from the max bandwidth TB4 can offer.

    What I said: "Correction: It will offer up to 80 Gb/s when in a single direction. 40 GB/s is for bi-directional use. So, for example, DisplayPort 2.0 will be using it at 80 Gb/s."

    Keep in mind that the 80 Gb/s is achieved via simplex link over a TB3 cable with the proper hardware. That's it and all I'm talking about, entirely independent from what controllers we might be thinking about or referring to in the article.
  • Deicidium369 - Thursday, January 14, 2021 - link

    DP2.0 has been talked about over TB4 - but haven't seen it in the wild yet. TB4 will allow 80Gb/s in single direction connections like DP2.0
  • KPOM - Wednesday, July 8, 2020 - link

    Will we be able to use a 2 meter Thunderbolt 4 cable with existing TB3 ports, and get the benefits of full USB 3.2 speeds? Right now, the only 2-meter TB3 cable I know of that also supports 10gbps USB 3.2 is Apple’s $129 cable.
  • Valantar - Wednesday, July 8, 2020 - link

    I was wondering the same thing - longer cables is one of the most basic needs for TB3. Of course if I/O speeds are to keep rising we'll need to move tho active cabling as a default sooner rather than later (with all the added cost that brings), but for now I would love 2m TB4 cables to also work as TB3 and USB 3.2 cables (given that ny 10Gbps USB-C cable should work as a TB3 cable this ought to be possible). Though i suspect what they are doing is making a low-cost active cable with limited length, which would limit its backwards compatibility.

    Limiting the 4-way hubs to what seems to be a minor update to a protocol also seems like an odd lock-in for no real reason. Couldn't support for this be patched into existing controllers?
  • repoman27 - Wednesday, July 8, 2020 - link

    The $129 Apple cable is active, hence the price tag. The problem with active cables that support multiple protocols is that you need redrivers / retimers that can handle all of the necessary signaling rates. For Thunderbolt 4 that would include:

    USB4 Gen 3 / Thunderbolt 4 / DisplayPort UHBR 20 - 20 Gbit/s
    USB4 Gen 2 / USB3 Gen 2 / DisplayPort UHBR 10 - 10 Gbit/s
    USB3 Gen 1 - 5 Gbit/s
    Thunderbolt 3 - 20.625 Gbit/s
    Thunderbolt / Thunderbolt 2 - 10.3125 Gbit/s
    DisplayPort UHBR 13.5 - 13.5 Gbit/s
    DisplayPort HBR3 - 8.1 Gbit/s
    DisplayPort HBR2 - 5.4 Gbit/s
    DisplayPort HBR - 2.7 Gbit/s
    DisplayPort RBR - 1.62 Gbit/s

    In either direction on each of the four high-speed differential signaling pairs. Even if someone starts cranking out suitable low-power chips that can handle this task, a 2.0 m cable is still probably going to cost somewhere in the neighborhood of $80.
  • repoman27 - Wednesday, July 8, 2020 - link

    Oh, and re: the hubs, Intel has never offered a discrete Thunderbolt controller with more than two ports. Being limited to one port for upstream and one port for downstream made daisy chaining the only possible topology. Now with Goshen Ridge, vendors can finally build hubs with a single upstream facing Thunderbolt port and up to 3 downstream facing Thunderbolt ports.

    And no, existing controllers can't be patched to add additional ports or USB4 support. However, Thunderbolt 4 docks / hubs will be compatible with Thunderbolt 3 hosts / devices.
  • Valantar - Wednesday, July 8, 2020 - link

    According to the spec table/compatibility list in the article the 4-port hubs ("accessories with four thunderbolt ports") are only compatible with Thunderbolt 4, and there is no mention anywhere about these being backwards compatible. Which, as I said, seems like an unnecessary lock-out which I am surprised if couldn't be patched in via firmware of they wanted to. I never said anything even remotely close to patching in support for more lanes from the controller, nor patching in TB4 support.

    To your previous comment, as I said, I expect Intel's new 2m cable to be a cost-down active cable, but a passive cable would be really great exactly because backwards compatibility ought to be simple.

    All in all, not to sound crass, but maybe work on your reading comprehension a bit? Both of your comments address things my comment already addressed...
  • repoman27 - Wednesday, July 8, 2020 - link

    No offense taken, and if you think my reading comprehension is off, you should check out my listening comprehension!

    Sorry, my first comment there was more directed at KPOM. And my responses reflected my reading of the specs and Intel's slide decks rather than what Ian wrote. At 20 Gbit/s, passive cables are limited to 0.8 m, and a 2.0 m cable will necessarily be active. I'm pretty sure Ian has mistaken Intel marketing's use of the royal "we" when it comes to the cables. That's really not Intel's bag. And my point was that if Apple is currently selling nearly the same cable for $129, even Monoprice will have trouble cracking $80. And if Intel were to start making the redriver / retimer chips for these types of cables, I wouldn't expect prices to go down at all. So to sum up, Intel doesn't have a 2m Thunderbolt 4 cable, but if they did, it would be both active and expensive.

    Also, by way of the USB4 spec, all of the downstream facing ports on Thunderbolt 4 docks and hubs as well as the upstream facing port on Thunderbolt 4 docks are required to support interoperability with Thunderbolt 3 products. Therefore the Goshen Ridge silicon necessarily supports Thunderbolt 3 on all 4 ports. Which also means it is fully backwards compatible with Thunderbolt 1 and 2 using an adapter.

    You said, "Limiting the 4-way hubs to what seems to be a minor update to a protocol also seems like an odd lock-in for no real reason. Couldn't support for this be patched into existing controllers?"

    I read that as, "Limiting the 4-way hubs to [Thunderbolt 4] seems like an odd lock-in for no real reason. Couldn't support for this be patched into existing [Thunderbolt 3] controllers?"

    Therefore my response was that the 4-port silicon really does support all the Thunderbolts, but obviously you can't add USB4 or more ports to the existing Thunderbolt 3 chips via a firmware update.
  • Jon Tseng - Wednesday, July 8, 2020 - link

    Dumbass question - will TB4 make an eGPU soln faster?

    Asking for a friend who might want to fire up Cyberpunk 2077 on his hypothetical Tiger Lake laptop..
  • Techtree101 - Wednesday, July 8, 2020 - link

    Soln? Potentially, yes, because more power can be fed to a eGPU solution.
  • Jon Tseng - Wednesday, July 8, 2020 - link

    Hmm. Was thinking more in terms of some external GPU being bottlenecked in terms of PCI lanes (or other bandwidth factors) when being used externally. I guess an RTX is gonna have an external PSU anyhow. Dunno if TB4 helps the situation re: bandwidth?
  • Eliadbu - Wednesday, July 8, 2020 - link

    eGPU solutions usually have their own power supply and they also deliver power to the computer. So I don't see this enabling them to be faster. The issue is the limited data rate of the PCI-E 3.0 4x in addition to other overheads that come due encapsulating the PCI-E when transmitting the through thunderbolt connection.
  • Mr Perfect - Wednesday, July 8, 2020 - link

    This is cool tech, so hopefully the OEMs actually put some TB4 ports on things. It's been about six years since USB-C rolled out and we're still getting between 0 and 1 ports of that.
  • tipoo - Wednesday, July 8, 2020 - link

    Why does it require twice the host bandwidth when the TB bandwidth is the same?
  • repoman27 - Wednesday, July 8, 2020 - link

    I believe the first comment in this thread pretty much answers you question—because OEMs sometimes used PCIe Gen 3 x2 links or only a single DisplayPort stream for Thunderbolt 3 controllers. So now with Thunderbolt 4, you need to provide each host controller a minimum of a PCIe Gen 3 x4 link and two DisplayPort main links capable of HBR2.

    31.5 Gbit/s of PCIe and 34.56 Gbit/s of DisplayPort data seems like a perfectly reasonable requirement for the back end of a controller that enables two downstream ports at 40 Gbit/s each.
  • FAdamsXII - Wednesday, July 8, 2020 - link

    What good is an open hardware standard that requires use of a proprietary closed "standard" to be able to use it? Thunderbolt 3 is touchy enough as it is, and while most of this seems like it'll reduce complexity, requiring use of a proprietary standard that IIRC is patented by Intel while pretending anyone can use it and its an open standard is typically Intel, and that's not meant as a compliment.

    They wonder why their semi-proprietary products like Optane don't really go anywhere, but parading around hardware that isn't open while pretending that it is might have a lot to do with that.
  • repoman27 - Thursday, July 9, 2020 - link

    While I'm with you on the feeling that this is only a veneer of openess and typically Intel, the USB4 specification is freely available and that more or less includes the Thunderbolt 3 protocol.

    All you need to do is sign the USB4 Adopters Agreement and you get access to a royalty-free reasonable and nondiscriminatory license from the Promoters and other Adopters to certain standard essential IP. You'll also need to sign the USB 3.0 and USB 2.0 Adopters Agreements, but those are also reciprocal, royalty-free licensing arrangements for compliant products.

    If you demonstrate compliance with the specs through the USB-IF testing program you are granted certain rights to use the USB-IF logos.

    If you complete certification and execute a Thunderbolt trademark license with Intel you can use the Thunderbolt trademark or logo on your product.

    You'll probably also need to become a VESA member to access the DisplayPort specs, and they do charge a membership fee based on annual sales revenue, however there are no royalties or licensing fees associated with their standards.

    So really it comes down to how much of a minefield developing your own USB4 / Thunderbolt IP ends up being. Honestly I don't expect anyone outside of Apple and maybe Asmedia or AMD to bother.
  • hubick - Thursday, July 9, 2020 - link

    I spent a lot of money on a 2TB Samsung X5 Thunderbolt 3 external SSD, which I can now no longer use on my new Threadripper system. Intel can shove any standard that doesn't have regular add-in cards that work on AMD systems. I won't fall for their proprietary crap again, no matter how fast it is.
  • Valantar - Friday, July 10, 2020 - link

    There are ways to make any TB3 AIC work on any system, IIRC you just need to short two pins on the connector for the cable that would connect to the motherboard. Last I read hot-swap was a bit iffy with this workaround, though that's several years ago.
  • Kamen Rider Blade - Thursday, July 9, 2020 - link

    Yet there still isn't a speed improved variant.

    Kill the last wires used for USB 2.0 backward compatibility and turn those into the 3rd set of Full Duplex lanes.

    It's been done as a Alt mode already by another group.

    Just make it official and kill off USB 2.0 support within the cable.

    Bandwidth is more important and TB4 / USB4 should be pushing for more speed.
  • ET - Thursday, July 9, 2020 - link

    I might be misunderstanding something, but wouldn't a USB 4 40 cable cover all the other standards as well?
  • vortmax2 - Thursday, July 9, 2020 - link

    "The first devices to come to market with TB4 support will be Intel’s Tiger Lake platform, for use in laptops, which will have TB4 baked right into the silicon."

    Does this mean ALL laptops with the Tiger Lake will actually have TB4 or is there other things manufacturer need to do for enablement?
  • Santoval - Friday, July 10, 2020 - link

    So Thunderbolt 4 is basically Thunderbolt 3+. Not a truly new version, just an optimization of the existing one, with some extras and better security. Maybe, just maybe, it should have been called either TB 3+ or TB 3.1 to point that out?
  • repoman27 - Saturday, July 11, 2020 - link

    One more time: Thunderbolt 4 fully supports USB4, uses it as the default signaling mode, and implements the Thunderbolt Alternate Mode to support backwards compatibility with Thunderbolt 3 and previous versions.

    Thunderbolt 3 does not support USB4 signaling at all. Therefore this is a fundamental change to the design and well worth incrementing the version number for. Maybe, just maybe, you're missing that point?
  • Santoval - Saturday, July 11, 2020 - link

    I am missing no point. Having Thunderbolt support USB4 (and previous), ensuring backwards compatibility and, er, simplifying(?) the huge cable mess/hole USB-IF dug themselves into, falls under the "some extras" & "optimization" category I mentioned above, and is in no way a "fundamental change" to the protocol. Besides, USB4 itself was based on TB3, even if the latter apparently does not support the signalling of the former, which is bizarre.

    When you trim, optimize, modify or extend an existing interconnect protocol (either internal or external) without upgrading it to higher speeds like what happened with *all* its previous versions it is not a truly new protocol by its *own* definition, not mine. TB4 is a modification, extension etc of TB3, not its actual successor, which is why it does not deserve a full new number revision.

    Just imagine PCI-SIG releasing PCIe 6.0 without doubling the speed over PCIe 5.0 but with adding a missing feature like full cache coherency and maybe making it compatible with CCIX and a couple of other interconnect protocols such as NVLink. Would this revision deserve a move to version 6.0 or would it more prudent to call it something like PCIe 5.1 (like they already do with minor revisions)? It's exactly the same with TB4.
  • repoman27 - Saturday, July 11, 2020 - link

    Except Thunderbolt 2 didn’t increase the signaling rate over the original version, it merely implemented channel bonding. Both are two channels at 10.3125 Gbit/s.

    So Intel’s choices were:

    1. Not support the new USB4 protocol and keep selling Thunderbolt 3 for the next few years until they can figure out how to deliver 40 Gbit/s PAM4 on a consumer I/O port.

    2. Support USB4 and Thunderbolt 3, which makes sense seeing as they have already integrated both into their CPUs and developed discrete controllers with similar capabilities including the first USB4 hub silicon, but simply bill it as “USB4 with Thunderbolt 3 Interoperability”.

    3. Support USB4 and Thunderbolt 3, set higher minimum performance targets and make certain aspects of those specifications normative rather than optional, develop a testing and certification program for hardware that also meets these stricter requirements, and refer to that as “Thunderbolt 4”
  • jabber - Saturday, July 11, 2020 - link

    Well managed to miss 1, 2 and 3 so far so maybe version 6?
  • KimGitz - Wednesday, July 15, 2020 - link

    Apple committed to continue supporting Thunderbolt and Intel baking it in their Tiger Lake, all I will really need to check when buying a computer is support for TB4.
    With a Thunderbolt 4 enabled computer I don't have to worry about whether the device is USB or Thunderbolt or what generation.
    One of the benefit of making Thunderbolt 4 also USB4 compliant is the new accessories architechure.
    Which makes me wonder if we can have a 4 port Dock connected to 4 Thunderbolt computers for 10G networking?
    If we look at the design wins for Ice Lake it is not difficult to imagine the kind of market presence TB4 will have once Tiger Lake comes out. I expect Microsoft will finally implement Thunderbolt on their next Tiger Lake powered Surface devices now that Intel has addressed security concerns.
    I was actually shocked to see Asrock with desktop motherboards supporting AMD Ryzen CPUs and Thunderbolt 3. Hopefully we will see more desktop motherboards supporting TB4.

    There is now a healthy selection of devices and accessories supporting TB3. Generally there is a performance increase between similar USB and Thunderbolt devices e.g storage, audio interfaces.

Log in

Don't have an account? Sign up now