Slashdot Mirror

← Back to Stories (view on slashdot.org)

Upcoming USB 3.2 Specification Will Double Data Rates Using Existing Cables (macrumors.com)

Posted by BeauHD on from the two-is-better-than-one dept.

A new USB specification has been introduced today by the USB 3.0 Promoter Group, which is comprised of Apple, HP, Intel, Microsoft, and other companies. The new USB 3.2 specification will replace the existing 3.1 specification and will double data rates to 20Gbps using new wires available if your device embraces the newest USB hardware. Mac Rumors reports: An incremental update, USB 3.2 is designed to define multi-lane operation for USB 3.2 hosts and devices. USB Type-C cables already support multi-lane operation, and with USB 3.2, hosts and devices can be created as multi-lane solutions, allowing for either two lanes of 5Gb/s or two lanes of 10Gb/s operation. With support for two lanes of 10Gb/s transfer speeds, performance is essentially doubled over existing USB-C cables. As an example, the USB Promoter Group says a USB 3.2 host connected to a USB 3.2 storage device will be capable of 2GB/sec data transfer performance over a USB-C cable certified for USB SuperSpeed 10Gb/s USB 3.1, while also remaining backwards compatible with earlier USB devices. Along with two-lane operation, USB 3.2 continues to use SuperSpeed USB layer data rates and encoding techniques and will introduce a minor update to hub specifications for seamless transitions between single and two-lane operation.

33 of 159 comments (clear)

  1. Drop the Serial by Big+Hairy+Ian · · Score: 5, Insightful

    Surely if they are going to have multiple data channels then this is a Universal Parallel Bus

    --

    Build a Man a Fire, and He'll Be Warm for a Day. Set a Man on Fire, and He'll Be Warm for the Rest of His Life.

    1. Re:Drop the Serial by iggymanz · · Score: 5, Funny

      next they'll add daisy chaining though it might need a terminator at the end and little DIP switches to set each device to a unique ID. It could be called something like Simultaneous Chained Serial Interfaces

    2. Re:Drop the Serial by Matt_Bennett · · Score: 5, Informative

      The difference is that the data on each line is sent serially, with embedded clocking and controlled skew between lanes- meaning that the data on each lane is somewhat independent. In general a parallel bus includes separate clocking. In a serial bus like this, the data is encoded in certain ways to allow the clock to be a part of the data- the most basic encoding is what is used in 10Mb Ethernet- Manchester, where every bit has it's own clock, and it goes up from there.

    3. Re:Drop the Serial by Anonymous Coward · · Score: 3, Funny

      If sacrificial goats become a requirement for proper operation, I'm going to pivot to a career in landscaping.

    4. Re:Drop the Serial by squiggleslash · · Score: 4, Informative

      It's harder than you think: as you increase the data rate, trying to synchronize all 8/16/32/whatever lines becomes close to impossible. Breaking the data into multibit chunks and sending each serially via a different channel is easier because each channel can run independently, without regard to latency.

      I believe this type of thing is also why we've moved away from, say, CPUs with direct access to memory (instead CPUs have multiple layers of cache between them and the computer's real memory.) It'd be nice and much more efficient to have the memory in your computer deliver up 64 bit words to the CPU at 4 gigawords a second (ie in sync with the CPU's 4GHz clock), but good luck trying to make a parallel motherboard bus that can deliver that.

      --
      You are not alone. This is not normal. None of this is normal.
    5. Re: Drop the Serial by Junta · · Score: 2

      Over RJ45, it got to 10Gbps (though there is also 2.5 Gbps and 5 Gbps, which are appealing for range and power requirement reasons, largely done due to the oddity that many wireless access points were constrained by their ethernet uplink, which is embarassing for etherenet...). Over SFP+, it's at 25 Gbps now, and QSFP is at 100 GBps.

      Of course, in terms of what's relevant to the sorts of systems that use usb seriously, it pretty much is at 1 Gpbs.

      --
      XML is like violence. If it doesn't solve the problem, use more.
    6. Re:Drop the Serial by KiloByte · · Score: 3, Funny

      If sacrificial goats become a requirement for proper operation, I'm going to pivot to a career in landscaping.

      It might be a complex science, but it is a science. Once you figure the correct type of dagger (both blade and handle), the number and color of required candles, you should be set.

      --
      The creatures outside looked from Alt-Right to Antifa; but already it was impossible to say which was which.
    7. Re:Drop the Serial by ctilsie242 · · Score: 2

      You sacrificed goats? I was lucky enough to just get away with chickens, especially when using differential SCSI.

    8. Re:Drop the Serial by AmiMoJo · · Score: 2

      Just to expand on that a little, the issue with parallel busses became that data on each line would arrive out of sync with other lines and the clock.

      For example, PCIe 4.0 is 16 gigabits/second per serial channel. Each bit has a width of 0.0625 nanoseconds, during which time light can travel about 18mm. If you wanted to transfer say 32 bits in parallel like the old PCI bus, you would need 32 connections. Problem is that for practical reasons the PCIe slot needs to be quite wide, so if your 32 pins are spaced over say 18mm then the extra distance that a signal at one end has to travel compared to a signal at the other end will mean its signal arrives one clock cycle later than the other.

      Then there is clock jitter. When you have a 16GHz clock, the variation from one cycle to the next is a huge problem and low jitter clocks at that speed are very expensive and tricky to use.

      You also need to use very small signals when you get up to those clock rates. There is something called slew rate, the speed at which voltage can change. You can either make the change happen faster, which we did and is now quite hard, or you can lower the voltage so it doesn't have as far to go. But lowering the voltage also makes it more prone to electrical noise.

      The solution is to use serial. One one differential signal for both clock and data. Differential signals have much better noise immunity, and as an added bonus you only need an 8GHz data rate on either line to get a combined data rate of 16Gb/sec. It also means your cables are smaller and easier to manage, the electronics are much simpler and cheaper etc.

      --
      const int one = 65536; (Silvermoon, Texture.cs)
      SJW, n: "Someone I don't like, and by the way I'm a fuckwit" - AC
    9. Re:Drop the Serial by snakeplissken · · Score: 2

      Once you figure the correct type of dagger (both blade and handle), the number and color of required candles, you should be set.

      I assume you have to have the proper dribbly wax?

      snake

  2. wrong direction by Build6 · · Score: 5, Insightful

    the "one plug for everything" trend that began with USB Type C is a step in the wrong direction.

    having "unique" plug types for particular purposes is a *feature*, not a bug - simply by looking at the plug, we know what the cable and the port does.

    Replacing all the legacy ports is necessary (if only because the old plugs are simply just too big for modern hardware), but replacing EVERYTHING with one plug, when everything now looks the same, you end up with a situation where you simply do not know if a cable or port can do what you need it to do.

    So you see a Type C plug - is it Thunderbolt or not? Is it a DisplayPort? What voltages/amps can it provide? nobody knows (where "nobody" can include the person whose hardware it is, much less someone else who has to work with it). just look at e.g. the many forum posts of people who connected "the wrong type" of USB-C-to-HDMI connectors because they didn't know their USB C wasn't the USB C that they thought USB C was supposed to be.

    this is made even worse considering that there's active circuitry involved, where you need to worry about whether the cable itself is built right (see e.g. Benson Leung's long list of cables that can fry your hardware). in the old days, a crap cable just means crap performance or no connection. not any more.

    1. Re:wrong direction by Viol8 · · Score: 5, Insightful

      "the "one plug for everything" trend that began with USB Type C is a step in the wrong direction."

      I disagree. I'm utterly fed up with multiple different types of USB plug, some of them very similar making it an utter pain finding a cable with the correct connectors on both ends. There is zero reason to have all these - its not for space reasons like RS232 had its 9 pin plug since even the largest USB connector is pretty small.

      "you end up with a situation where you simply do not know if a cable or port can do what you need it to do."

      In 99% of cases you simply need to connect 2 devices, its not complex. If there is a max voltage/current issue then colour code the cables, but DONT create yet another sphagetti soup of connectors.

    2. Re:wrong direction by thegarbz · · Score: 4, Informative

      No it's the right direction providing they get to the correct end goal. We're in a transition period where the capability of the spec is being built up. The end goal is perfectly in line with what you are saying: one plug, do everything, no confusion.

      It's just not there yet.

    3. Re:wrong direction by Build6 · · Score: 2

      Exactly. There will always be "optional extensions". You can never know visually (let's say there's any rule in colour coding to indicate capability, for example, well, can you trust the manufacturer put in the right colour?)

      That means that upon encountering a Type C cable, you can never know what it actually does (except some lower baseline of capability which may or may not be sufficient for you).

      This really matters because of the problem of time.

      A cable manufactured today cannot be expected to support standards that may crop up somewhere down the line, but that cable - if manufactured well, and we're not going to start proposing time-limited self-destructing cables are we? - is going to be around well past the time when whatever spec it was built to support was "complete".

      Even if "damage" is no longer a problem with some future improvements in self-protection in ports, we won't know capability visually, so are we supposed to sit down and plug in cables one by one to find out what they really do (remember - mere labelling cannot be trusted)? as opposed to the old days, when I can immediately know that a cable is a video cable because it won't fit otherwise

    4. Re:wrong direction by Build6 · · Score: 4, Insightful

      I think I need to point out something here - there's a distinction between "nice to have" and "practically unfeasible".

      there is a functional issue I think you're overlooking - you're complaining about things not fitting together with all sorts of different plug types (microUSB, mini-USB etc.) - things "should" be electrically and functionality/feature-compatible, but which you cannot connect together, therefore you're unhappy about it.

      You're actually looking at a subset of the entire space, because remember: in the real world, not everything is supposed to plug into everything else.

      It's not like if I can plug my handphone into my lawnmower it will automagically allow me to remotely mow my lawn from the office (actually, wait... somebody kickstarter this, quick). Making them plug-compatible achieves... what?

      Not everything can work together with everything else. that's the point of having different plug types

      having a "universal" plug type is something that would be VERY nice, just as it would be nice for unicorns to show up and make it rain cake (hey, they're my conception of unicorns so their powers are what I want it to be), but as a real world issue, you cannot build in sufficient forwards compatibility to make up for whatever may come up in the future, so even a "do everything" cable today will become "insufficient" in the future, and you're back to square one. Perhaps you can guarantee that every cable ever made now will always be useful in the future for at least plugging in your mouse and keyboard, but I really doubt there will be a shortage of mouse and keyboard cables in the future - the problem with Type C isn't at the "lowest common denominator" side of the market

      again, not everything is supposed to plug into everything else, at least not at today's technology level (and arguably this will never be appropriate).

      Let's say that in the future all lawnmowers use USB Type C plugs for power (so you could use the same cable to power your lawnmower as you use to charge your laptop). It then becomes physically possible to plug anything with a USB Type C port into your lawnmower.

      Does that achieve anything if they simply don't do the same things?

      Does plugging in your video projector into your lawnmower mean that you can have video recordings of you mowing the grass projected out onto a screen? Maybe, maybe not - let's say there's a niche demand for this, well, if video output was something your super duper special lawnmower is supposed to do, then having a VGA port, say, on your lawnmower makes sense, and you can tell video output is something it can do. However, your neighbour's lawnmower, which doesn't do video output, won't need that. But if they have a Type C port... is everything in the future to have stencilled/printed on the underside a full list of capabilities? Is that what we're supposed to do? I don't understand why it's such a problem that, if two things aren't supposed to work together, their ports physically don't match.

      The problem now with a Type C port is, what it REALLY says, is that it MAY OR MAY NOT work with something else that has a Type C port. That's NOT an improvement.

    5. Re:wrong direction by Matt_Bennett · · Score: 2

      Claiming to and actually making a cable that can still meet the spec after the specified number of insertions and removals is 2 different things. If someone is claiming this and doesn't have a traceable USB logo (back to the USB-IF) I wouldn't count on it.

    6. Re:wrong direction by ColdWetDog · · Score: 2

      Maybe you're channeling Mr. Ive?

      Three plugs for the Mac Mini under the desk,
      Seven for the ancient X-serve in their halls of wire,
      Nine for MacPros doomed to die,
      One for Jony Ive on his dark (but tastefully brushed aluminum) throne
      In Cupertino where the Shadows lie.
      One Plug to rule them all, One Plug to find them,
      One Plug to bring them all and in the darkness bind them
      In Cupertino where the Shadows lie.

      --
      Faster! Faster! Faster would be better!
    7. Re:wrong direction by Ramze · · Score: 2

      I'm waiting to see how this all plays out before I make any judgments, and I'm curious how devices are going to identify and select the proper use for the same USB cable. For instance, I'm used to having USB for I/O and HDMI for one-way A/V. If I connect 2 devices that both previously had HDMI out ports AND HDMI in ports (for video capture) as well as USB ports for I/O, how will they know what I want them to do if there's a single port on both and I connect them together? Will they both think I want to open a file from the other device? Will they both try to cast video and audio to the other device or expect A/V back or are we relying on software to negotiate what gets cast to what?

      Will there be separate USB C ports for each use, (A/V in, A/V out, Data I/O) and if so, how will I be able to tell them apart on an unfamiliar device?

      It seems separate ports would be wise -- one for power, one for A/V import, another for A/V export.... possibly another for data as that can set a device into a different mode when plugged in. I dunno. *shrugs*

      I'm glad that I might be able to cut down my cable collection, but I'm not sure yet how we're going to use the same shape port and same type cable for so many things -- especially combining typically one-directional A/V communications with standard bi-directional USB communications between devices.

  3. That's easy enough to do by DrXym · · Score: 5, Funny

    Just stop transmitting the 0s. The data rate is instantly doubled.

    1. Re: That's easy enough to do by qbast · · Score: 4, Funny

      And when we stop sending 0s, then we know exactly what will be sent - only 1s, so stop sending them as well. Instant infinite bandwidth and no cable is required!

    2. Re:That's easy enough to do by MysteriousPreacher · · Score: 4, Funny

      It depends. If the cable is angled downwards then the 1s, being heavier, fall more quickly than 0s. If the cable is angled upwards then the 0s are lighter and move more quickly.

      So whether you encode your data in to 0s or 1s will depend on how the cable is angled.

      I think we're wasted here. We could be working for these guys: http://www.audioquest.com/ethe...

      To quote their 'tech blurb' for what is an 8m Ethernet cable, ordinarily retailing for less than $10, being sold for $1,158.75:

      DIRECTIONALITY: All audio cables are directional. The correct direction is determined by listening to every batch of metal conductors used in every AudioQuest audio cable. Arrows are clearly marked on the connectors to ensure superior sound quality. For best results have the arrow pointing in the direction of the flow of music. For example, NAS to Router, Router to Network Player.

      --
      -- Using the preview button since 2005
    3. Re: That's easy enough to do by Soleen · · Score: 2

      I disagree, you still need to know how many 1s need to be sent. So, not sending 0s saves half of the bandwidth, but we still must know how many 1s we want to send. So, unfortunately, we can't avoid cables yet. However, we can compress: for example, we want to send 64 of 1s: which is 1000000b in binary, thus 64 can be described only by one 1. So, we can send one 1 instead!

      --
      LiFe iS bEAuTiFul :-)
  4. Re:Typo? by enriquevagu · · Score: 2

    No, 2 GB/sec is 16 Gb/sec, and this typically refers to effective transfer values (discounting overheads) so the speed is roughly double (ermm, sort of).

  5. Cables vs wires by enriquevagu · · Score: 2

    The post is quite confusing, since the headline says Double Data Rates Using Existing Cables but the summary specifies double data rates to 20Gbps using new wires.

    The explanation is that USB-C connectors (type-C connectors) have two different sets of transmission/reception pairs, two pairs for transmission and two pairs for reception. This differs from previous type-A and type-B connectors, which only have one pair of pins for (bidirectional) data transmission.

    All type-C connectors implement the pins for the four pairs, but many cables only populate the wires for two pairs, one for transmission and one for reception (as clearly explained in the Wikipedia link provided). Given the simmetry of the connector, I bet this was designed to support the reversible connection, in a manner in which only one of the pairs works at a time. For this reason, the new USB 3.2 might double the speed, but it would require cables that implement the complete set of wires for such speed. So, in the general case, double speed would actually require new cables with the complete set of wires.

    1. Re:Cables vs wires by msauve · · Score: 2

      "All type-C connectors implement the pins for the four pairs, but many cables only populate the wires for two pairs"

      Nope. If a cable only implements 2 pairs, it's not a USB cable regardless of what the seller calls it. The spec requires cables to implement all 4 pairs.

      More correctly: There are Type-C to other type cables which only have 2 pair, simply because it's not possible to have a 2nd pair on the other end (e.g. Type-C to USB 3.1 Standard-A). There's also an uncommon USB 2.0 Type C cable, which doesn't have any of the high speed pairs. But there is no C to C cable allowed which only has 2 high speed pairs.

      Source: Universal Serial Bus Type-C Cable and Connector Specification (the requirement has existed since the original 1.0 spec)

      --
      "National Security is the chief cause of national insecurity." - Celine's First Law
  6. I look forward too... by trabby · · Score: 2

    I look forward to USB 3.0 and 3.1 becoming "USB 3.2 Gen 1", "USB 3.2 Gen 2" and actual USB 3.2 being "USB 3.2 Gen 3".

    Seriously these marketing idiots need to be shot.

  7. Common connectors are a great thing by sjbe · · Score: 2

    the "one plug for everything" trend that began with USB Type C is a step in the wrong direction.

    Could not disagree more. There are HUGE advantages to having common connectors. These advantages vastly outweigh the drawbacks. Connectors should be commonized as much as possible. The fewer number of cables types I have to deal with the better. I basically want to be able to hook up nearly everything with 1 or at most 2 types of cables.

    having "unique" plug types for particular purposes is a *feature*, not a bug - simply by looking at the plug, we know what the cable and the port does.

    Except you don't and you never did in a great many cases. Having to carry around and deal with 20 different types of cables is wasteful and unnecessary.

    So you see a Type C plug - is it Thunderbolt or not? Is it a DisplayPort? What voltages/amps can it provide?

    All good quality USB-C cables will work for Thunderbolt. Same with Displayport. As long as you are using good quality cables it is a non-issue. Sourcing good quality cables is not a difficult problem.

    this is made even worse considering that there's active circuitry involved, where you need to worry about whether the cable itself is built right (see e.g. Benson Leung's long list of cables that can fry your hardware)

    If you buy a crap cable from a crap vendor be prepared to get crap results. This is nothing new and has nothing to do with whether or not common connectors are a good idea or not. While I do generally agree with the principle that cables should be dumb and the smarts should be in the devices it's not something I'm going to make a holy war over if it gets the job done.

    1. Re:Common connectors are a great thing by Junta · · Score: 2

      If you buy a crap cable from a crap vendor be prepared to get crap results.

      Of course at the amperage of typical usb connection before, the 'crap results' were 'the damn cable didn't work, what a waste'. With USB power delivery, the 'crap results' are 'my thousand dollar laptop fried'. We have the unfortunate circumstance of 'cheap power cable' vendors, 'high current', and 'small form factor that mfg can screw up'.

      Same *could* happen with C13 power cabling in theory, but those are so gigantic it's hard to screw up, or at least screwing up is not a natural consequence of trying to cut costs.

      --
      XML is like violence. If it doesn't solve the problem, use more.
  8. Too confusing by blindseer · · Score: 4, Interesting

    I have to wonder what the USB people are thinking. They have a good idea going here, why are they trying to ruin it?

    It used to be that USB-C was USB but faster. It made the "On-The-Go" bi-directional features from USB2 required so people didn't have to worry about the host/device relationship too much. The Micro-AB connector was switched to a "flippable" version, they called "Type C", which is great. The new connector is just as small but handles more power, and I don't have to worry if I got the "right" end or which way is "up".

    Where they are starting to mess this up is with four, FOUR, different video modes. You have Thunderbolt (which is just DisplayPort mixed with PCIe data), DisplayPort (which may or may not be backward compatible with the Thunderbolt mode), HDMI, and MHL (which may or may not be backward compatible with HDMI). They should have told the HDMI people to piss off and stopped at three. Maybe even tell the MHL people to piss off too but they already had an agreement with MHL on the micro-USB connector.

    Now you have four video modes, two data modes (ThunderBolt and "SuperSpeed" USB), and with this new "Ludicrous Speed" mode they added a third data mode. It's already confusing on what video adapter or cable a person might need. Even buying a simple data cable is confusing. There's the USB2 cable, the USB3 cable, and the ThunderBolt cable, they all look identical at a glance with USB-C connectors on both ends. Will my expensive Thunderbolt cable support this new USB 3.2 data rate? Will it fall back to USB 3.1 speed nicely? Or will it crap out and support only USB2 speed?

    They created this "SuperSpeed" naming to differentiate the USB2 speed devices from the faster USB3 ones. Then when USB-C came along with two "SuperSpeed" lanes they had to figure out how to make that clear to the user. They came up with "SuperSpeed+". (Notice the addition of the plus sign? It's easy to miss.) What is this new one going to be called?

    I have to wonder if this is going to die before it even gets started. The people that want a faster USB got ThunderBolt already.

    --
    I am armed because I am free. I am free because I am armed.
    1. Re:Too confusing by Kjella · · Score: 2

      Where they are starting to mess this up is with four, FOUR, different video modes.

      Not really. Thunderbolt is external PCIe with DisplayPort as an alternate video mode, it's an alternative to USB data signaling not video. MHL is essentially HDMI squished to use fewer connectors + power supply, if you got USB-C out you have enough connectors for full HDMI and all MHL receptors are also HDMI ports. The problem is that if you use HDMI mode you don't get the power supply, the only way to get that from a USB-C device connected to a MHL-enabled TV is to fake MHL. That problem goes away if you got a USB-C receptor that can talk USBs power delivery standard.

      So really you have:
      USB data - suitable for "devices" like printers, scanners, headsets, keyboard, mouse etc.
      PCIe data - suitable for "extensions" like external GPUs, storage arrays etc. that could have been on the inside.
      DisplayPort video
      HDMI video
      Bi-directional power delivery with negotiation of capabilities

      Thunderbolt = PCIe data + DisplayPort
      MHL = HDMI(-ish) + power

      The only real redundancy there that is not there for compatibility reasons is DisplayPort and HDMI. As long as TVs mainly come with HDMI and monitors mainly with DisplayPort with no clear winner in sight that's probably not too stupid. In fact PCIe, DP and HDMI are now bundled in TB3 so in practice you have only two ports:

      TB3: All of the above
      USB3: USB data + power

      using one connector, USB-C. You still have five support levels though, USB speed level, PCIe speed level, DP level, HDMI level, power level. It's hard to be everything for everyone...

      --
      Live today, because you never know what tomorrow brings
  9. This is a solved problem by Black.Shuck · · Score: 2

    Cable-markers my good fellow.

    If you have enough cables to worry about what goes where, then cable-markers will very likely be part of your inventory.

    Or at the very least you'll have cables of different colours.

  10. The new keyboards by Provocateur · · Score: 2

    The new keyboards that will have USB can now be equipped with cigarette lighters.

    --
    WARNING: Smartphones have side effects--most of them undocumented.
  11. Too many chnages in too short of a time... by QuietLagoon · · Score: 2

    USB is becoming a big mess of different versions, far too many versions. USB is starting to lose its usefulness. USB 3.2? Most of my devices don't even use USB 3.0 yet. The USB spec is starting to look like changes are being made for the purpose of making changes, but to no real end.