Catch up on stories from the past week (and beyond) at the Slashdot story archive

 



Forgot your password?
typodupeerror
×
IT Technology

USB-C Can Hit 120Gbps With Newly Published USB4 Version 2.0 Spec (arstechnica.com) 69

An anonymous reader shares a report: We've said it before, and we'll say it again: USB-C is confusing. A USB-C port or cable can support a range of speeds, power capabilities, and other features, depending on the specification used. Today, USB-C can support various data transfer rates, from 0.48Gbps (USB 2.0) all the way to 40Gbps (USB4, Thunderbolt 3, and Thunderbolt 4). Things are only about to intensify, as today the USB Implementers Forum (USB-IF) published the USB4 Version 2.0 spec. It adds optional support for 80Gbps bidirectional bandwidth as well as the optional ability to send or receive data at up to 120Gbps.

The USB-IF first gave us word of USB4 Version 2.0 in September, saying it would support a data transfer rate of up to 80Gbps in either direction (40Gbps per lane, four lanes total), thanks to a new physical layer architecture (PHY) based on PAM-3 signal encoding. For what it's worth, Intel also demoed Thunderbolt at 80Gbps but hasn't released an official spec yet. USB4 Version 2.0 offers a nice potential bump over the original USB4 spec, which introduced optional support for 40Gbps operation. You just have to be sure to check the spec sheets to know what sort of performance you're getting. Once USB4 Version 2.0 products come out, you'll be able to hit 80Gbps with USB-C passive cables that currently operate at 40Gbps, but you'll have to buy a new cable if you want a longer, active 80Gbps.

This discussion has been archived. No new comments can be posted.

USB-C Can Hit 120Gbps With Newly Published USB4 Version 2.0 Spec

Comments Filter:
  • No, it can't. (Score:2, Informative)

    by Anonymous Coward

    Theoretically, sure. But it's USB. If you get 60% of rated it's more than you can realistically expect.

  • by nightflameauto ( 6607976 ) on Wednesday October 19, 2022 @12:02PM (#62980517)

    The way USB conventions for naming are going, it won't be long before we see:

    ANNOUNCEMENT: USB6.0 version12, subversionC, revision8, deltaB, lanecall4345231.
    Warning: Compatible but not fully compliant with similar but not same USB6.0 version12, subversionC, revision8, deltaB, lanecall4345230.

    • They'll run out of colors that are distinguishable before too long https://ourtechroom.com/tech/g... [ourtechroom.com]
      • Nobody really adheres to the color standards now anyway. And WTF is HP's Superspeed ports? Seems like that's a random smattering of various USB "standards" that's different even across the same model year, depending on which system/board you get.

      • by dargaud ( 518470 )
        I'd never known USB could have colors. I just checkes several laptops and towers I have handy, and they are all black. Even though some have the SS label, others the power label, others both and others none.
        • If you look at some USB 3.x connectors you will see that early USB 3 ports and cables were colored blue and some USB with PowerDelevery were red. I know the Raspberry Pi 4 boards use blue on the USB 3 connectors. https://cdn-shop.adafruit.com/... [adafruit.com]

          Other manufacturers are using the SS label instead, but blue connectors seem to be on some laptops.

    • The nice thing about standards is that you have so many to choose from; furthermore, if you do not like any of them, you can just wait for next year's model." -- Andrew Tanenbaum
    • Hush! Don't give them any ideas
    • So what? It's much better than having hundreds of different connectors because a faster connectivity standard is made.

      It will be just like WiFi, slow/old 802.11b devices gradually get phased out, and new 5-6 GHz devices arrived. Various revisions of the 802.11 standard increased speed while remaining mostly backward compatible.

      It's the same thing with USB. The USB-C connector is there for a while, and they will increase throughput and power. Does it mean every device and cable will benefit from all capabili

      • by aitikin ( 909209 )

        So what? It's much better than having hundreds of different connectors because a faster connectivity standard is made.

        It will be just like WiFi, slow/old 802.11b devices gradually get phased out, and new 5-6 GHz devices arrived. Various revisions of the 802.11 standard increased speed while remaining mostly backward compatible.

        It's the same thing with USB. The USB-C connector is there for a while, and they will increase throughput and power. Does it mean every device and cable will benefit from all capabilities of USB 6.0? No, but who cares?

        So let's make Thunderbolt use the C connector as well. But not make USB encapsulate Thunderbolt, so that now, consumers may think they have Thunderbolt because the port they have uses the same connector, spend thousands of dollars on the product that needs Thunderbolt, and not be able to use it because they actually only have USB3C...

        And before someone tries to point out that USB4 can encapsulate Thunderbolt, please remember that the USB-IF, in their InFinite wisdom, decided that it was not mandatory for U

      • I was more making fun of the naming conventions getting ever more ridiculous than the fact that the "standard" is absolutely not a standard in any real sense of the word. Pardon me.

        • Well of course I agree the naming is ridiculous (why not USB 4.1?) but it doesn't really matter for the end user, which will just use the same USB-C port the same way they always did. At some point down the road, they might get a device which will communicate at 120 Gbps with their PC, and the user won't care how it's called.

    • by dohzer ( 867770 )

      Why not go full-videogame on people and call the next generation USB1. Or even just USB?!

      • USBX!

        The 'X' is, of course for, XXXXXXXXXXXXXXXXXXXXXXXXTREEEEEEEEEEEEEEEEEEEEMEEEEEEEEEEEEEE!

        To be superseded by USBX 2.0 within one month.

  • by brunes69 ( 86786 ) <slashdot@nOSpam.keirstead.org> on Wednesday October 19, 2022 @12:14PM (#62980541)

    People harp on USB all the time, but let's remember that good-ol' ethernet started out only supporting 3Mbps, then went to 10Mbps, 100Mbps, 1Gbps, 10Gbps, and now we're all the way up to 400Gbps. Every time new speeds came out, new requirements were put on the physical layer. You can't just take any random Ethernet cable and expect 10Gbps speeds, there are many different classes of cable. It is really no different from USB.

    • If you think about it, USB is just "short ethernet" anyway. Even packetizes as TCP/IP last time i looked at my USB bus in wireshark
      • As someone that has designed with Ethernet and USB (at hardware and software levels), USB and Ethernet are VERY different (A canal and a railroad share a lot of characteristics, but, oh god, the differences!). The USB model is very much a host/device type network, point-point, and Ethernet is a peer-peer setup, usually physically point-point, but logically point-point or point-multipoint. USB is packetized, but looks like Ethernet to analyzers like Wireshark more for the convenience of Wireshark developers

    • by Anonymous Coward

      I remember the joys of 10BASE2 and loose crimps, thanks.

      Ethernet is another example of over-promising and under-performing technology, capturing the market and leaving technically better but more expensive solutions in the dust. For golly, it's cheap, and it does (pretends to do) the same thing therefore it must be just as good as the more expensive kit, no? At least it plugs in easy! This is a common theme in computing.

      Though they've been at it for a while longer than USB. So I say your comparison is a t

    • A consumer of tech equipment, like a smartphone, tablet, laptop, should have the right not to be technical, and to have things out of the box just work, based on common sense assumptions.

      Assumptions like: Oh great, a USB-C cable. Let me plug it in to this USB-C port. Great that I don't need to know which way is up! Wait! Why isn't this working? Q@#$% !!

      If they're going to significantly change (upgrade the capabilities) of the communication and power cord, they should either change the port, so it's obvious
    • Question- Is the USB spec changing faster and/or more often than the ethernet spec?

      When I first started getting involved with PC networking 100mbits what ya had at work and 10 is what you had at home. After some time had passed setting up 100mbits at home was something one could affordably do with a gigabit on the horizon. Fast forward another length of time and I've got a gigabit at home and ... hell I don't even know what we have at work anymore. The problem is I don't actually recall the dates with

      • Ethernet started going downwards with 2.5 and 5 Gbps as well as 10 Gbps. And you still get 1 gigabit on most (regular/cheap) motherboards and computers.

        USB started well with USB1 at 12 Mbps, USB2 at 480 and USB3 at 5 Gbps. At least 10x faster on every generation. But since then, it's mostly 2x increases and decimal revisions such as USB3.1, 3.2. USB4 didn't really deserve it's 4 (could have been 3.3) but the integration of thunderbolt was seen as an important step, I guess.

      • For day-day usage (not in datacenter), wired, unshielded twisted pair ethernet has progressed REALLY slowly. Spec wise, I think it was intended to go at a 5-ish year interval. The gigabit ethernet spec (1000Base-T) was ratified in *1999*, 10GBit in 2006 (but it was short range), 2.5GBit and 5Gbit in 2016.

        Really fast progress has been pushed to fiber, datacenter and multi-kilometer level runs. The physics of copper wiring/interconnect really limit the choices. Not much good in creating a consumer network whe

        • Which brings us to... will I reasonably be able to do peer to peer networking over this super duper fast USB cable? Because a lot of people only have a couple of computers. Just being able to occasionally hook the laptop up to the desktop to dump files to an array would be neat. (Or directly to a NAS, obviously.)

          • Probably not (except if you are connecting your super duper fast USB to a super duper fast network adapter). USB is a host to device protocol. It does not play well with multi-host arrangements. I see nothing directly in the USB standard that would allow this. *MAYBE* if you used a USB C to USB C, and one side negotiates as host and the other side as a device, which gets into the rat hole of non-standard dual role protocols or ... shiver OTG.

            If peer-peer networking is what you really want, use a network in

            • That's too bad, firewire would do it, I hoped USB would get there eventually. I don't recall any firewire roadmap that had anything over 3.2Gbps though

      • Whereas it feels like the USB spec is changing every 18-months.

        USB 4.0 was released in 2019 (and is still basically no where to be found)
        USB 3.2 was released in 2017
        USB 3.1 was released in 2013.

        It's really not as bad as you make it out. What changes a lot are the substandards, especially USB-PD which gets revised more often than every 18 months, and the USB-C connector which gets updated with each major change to USB-PD spec.

  • need to split cable max speed and cable max power.

    As an cable for high speed networking may not need to be able to change an laptop.

    • need to split cable max speed and cable max power.

      As an cable for high speed networking may not need to be able to change an laptop.

      This is a path to madness. All the cables really need to do all the things, despite the fact that doing so drives the cost up. Having 8 different icons for what a cable can and can't do would be insanity. Doubly so since no manufacturer would do it right.

      • All the cables really need to do all the things, despite the fact that doing so drives the cost up.

        No they really don't. Not all cables are design to be the same for a reason. They differ in physical properties, not just cost. Not every application needs or wants a thick arse cable.

    • They do. Cable max speed is only tied to higher power for long cables. To get the max speed (80Gbps) out of a longer cable, the cable has to have active (powered) electronics in it- at both ends. High speeds require more power. To support a long cable without the electronics requires more expensive transceivers at both host and device. The idea, as I get it, most applications use shorter cables. The few applications that need long, high speed, cables, should be able to take the burden of the more expensive

  • by 0xG ( 712423 )
    I prefer something that works.
    • The nice thing about RS-232 is that here is over a billion different way to configure it, an only one of those configurations actually talks to the device you want to communicate with!
    • RS-422 is better, less noise, clearer signals but I hear there's this new RS-485 standard that will be approved any day now.

    • by leptons ( 891340 )
      I'm ditching USB for external storage media. It's failed me far too many times. The same drives when used in a SATA setup never fail, but USB always causes me problems, eventually, without exception. It's just flakey. I don't care where the problem is,
    • So you pick a standard infamously picky about the cable, signalling, speed, and type that leaves you endlessly guessing the connection parameters until some magic combination works?

      WUT?

  • Can it hit 120 Gbps over the cheap USB-C cable I bought at the Dollar Store? I'm thinking somebody needs to certify cables for higher speed operation.
    • Just because all USB-C is a connector. It's no different than having a cable with USB-A on one end and USB-B on the other. You needed to know if the cable was rated for 1.1, 2.0 or 3.0.

      A lot of USB-C cables date back to the first era, so they were all the same. Any USB-C cord would work with USB 3.1 or lower. It's really only cables designed for 3.2 and later that need differentiation. And USB-IF does have terrible logos a manufacturer can use that is barely readable at the size they'll be printed that

      • That what I love about Slashdot; people will go out of their way to give well thought out answers to even my stupidest questions! I agree with you on the labeling for the different USB standards being completely illegible, especially for someone with vision as poor as mine. I usually can't even figure out which USB-C port on a laptop can be used for the AC adapter.
  • by Virtucon ( 127420 ) on Wednesday October 19, 2022 @12:24PM (#62980581)

    DH - We're gonna have to go right to Ludicrous Speed!
    CS - Ludicrous Speed? Sir we've never gone that fast before! I don't know if the ship can take it!
    DH - What's the matter Colonel Sanders? Chicken?

    - Spaceballs

    I'm expecting USB - C cords to now turn plaid

  • I don't think too many people care about the speed. Sure, if you're getting an external drive, you'll want to make sure that the cable and port support the drive's speed so that you get the performance you're paying for, but for most devices, it just doesn't matter. For many of us, the fastest thing over USB-C that we use is ethernet at 1Gbps, so we're fine now.

    What does matter is whether it works. If I plug in a USB-C monitor, does the port support that? Will all the features of my USB-C docking statio

    • Comment removed based on user account deletion
      • by crow ( 16139 )

        I'm seeing some 2.5Gbps ethernet now. Having WiFi standards that at least theoretically exceed 1Gbps is going to be pushing this as much as anything. I'm guessing the 2.5 and 5.0 Gbps parts are cheaper, so it may be a while before 10 becomes standard unfortunately.

        • by leptons ( 891340 )
          Just go look around ebay. You can find all kinds of 10gbit equipment for bargain prices. Datacenters have to keep upgrading, and so ebay typically gets the stuff that's getting phased-out, at good prices. I set up a 40gbit Infiniband link between my two computers for about $100 total, and that was quite a few years ago.
    • by darkain ( 749283 ) on Wednesday October 19, 2022 @01:03PM (#62980701) Homepage

      I find it interesting that you say the fastest thing anyone connects to USB-C connectors is a 1gbps NIC, and then go on to talk about hooking up monitors.

      Those monitors are what need these massive amounts of bandwidth.

      For instance, I'm currently running a pair of 1440p/120Hz monitors from a single USB-C port. Somehow, I think that is significantly more bandwidth than the 1gbps NIC I have right now.

      Also, before the "nobody needs 120Hz" business, yes, I seriously notice the difference between 60Hz and 120Hz and it helps with the work I do. And also, 1440p/120 takes the same bandwidth as 4k/60Hz.

      Now that we're starting to push 5k display, 8k displays, and on the bleeding edge, 16k displays! That USB-C bandwidth becomes more and more important. This is also why they're pushing from 80gpbs bi-directional to 120gbps uni-directional, because displays don't need to return that much data back to the source.

    • Not true, with single connection to docking port I'm pushing two monitors with many gigabit/sec, gig ethernet, mouse, keyboard

      speed matters!

    • For many of us, the fastest thing over USB-C that we use is ethernet at 1Gbps, so we're fine now.

      Huh? I've seen more people use e-GPUs over Thunderbolt than I have seen ethernet. But more importantly USB is quite limiting for those using docking stations where a single cable is supposed to provide power, peripherals, and Displayport to a 4K monitor. Good luck then plugging even a fast USB stick into the monitor's hub and getting any decent speed.

      For what USB is currently intended for, it still has a major speed bottleneck.

  • I might get excited. Think of all those watts I can use to.. charge my phoney?
  • Is there classic rectangular version of USB-C ? Or is it like VGA vs HDMI port?

    Types of Video Ports WTF:
    It is important to note that you can plug a USB-C-only device into a Thunderbolt v3 or v4 port of a computer and it will work fine.
    However, plugging a Thunderbolt v3 or v4 device into a USB-C port of a computer, will not work unless the USB-C port explicitly specifies it supports Thunderbolt devices.

    https://www.meridianoutpost.co... [meridianoutpost.com]
    • USB-C refers to the physical connection and cable itself, nothing more.

      The non-reversable, square connector is USB-A. Which comes in flavors like 'normal', micro, mini.

  • USB-IF (Score:5, Funny)

    by bustinbrains ( 6800166 ) on Wednesday October 19, 2022 @02:49PM (#62981015)

    USB-THEN, USB-ELSE, USB-ENDIF

    Soon the consumer will be able to write Turing complete programs using USB cables at home!

  • ...because while USB C is touted as the 'universal plug for everything' my Dell laptop that ONLY charges by USB C has died because it (over one weekend) simply..."forgot" how to understand to take in power.

    3 more of our company laptops, (different model, bought this year) came with USBC power cords, all have died and required a mobo replacement to be functional again.
    In their cases (not mine) they actually have the OPTION for a barrel power plug, but default is USB-C. I have 5 more laptops of that model in

    • Sure the charger supplying power isn't sending more than the laptop requests? If so, blown safety fuse.
    • USB C is shit.

      No. Your suppliers is shit. There's no reason for USB PD chips fail at a rate any differently than any other device. And yes I too have had a Dell laptop just say f-you I'm not charging ever again. And it wasn't even using USB-C.

      Hardware faults happen.

    • You have a Dell problem not a usb problem.

No spitting on the Bus! Thank you, The Mgt.

Working...