USB 4 Will Support Thunderbolt and Double the Speed of USB 3.2 (engadget.com) 165
At a Taipei event earlier today, Intel revealed that USB 4 will once again utilize dual channels to achieve 40Gbps speeds, even on existing 40Gbps-certified USB-C cables. A report adds: Better yet, thanks to Intel finally offering Thunderbolt 3 to manufacturers with open licensing, USB 4 will be integrating this tech and thus effectively becoming the "new" Thunderbolt 3. In other words, USB 4 will pretty much be the mother of all wired connectivity options, and will be ready for more powerful PCIe plus DisplayPort devices. It is expected to take 18 months between the final spec of USB 4 being published in the second half of this year, and the first devices hitting the market, so don't expect to see USB 4-powered commercial devices until sometime in 2021.
Further reading, from last week: USB-IF Confusingly Merges USB 3.0 and USB 3.1 Under New USB 3.2 Branding.
Further reading, from last week: USB-IF Confusingly Merges USB 3.0 and USB 3.1 Under New USB 3.2 Branding.
Will the wires catch on fire? (Score:5, Interesting)
I Know, I know faster transmission, of data doesn't necessarily require more power. Just a higher frequency signals of the data. Which gets increasingly harder to read, and more prone to interference. But 40Gbs in a cable that most people will coil up to keep the wires organized just seems like something prone to problems. Unless USB4 cables will have a ton of insulation, to prevent the outside world from interfering with it. Or will it have more error checking thus this 40Gbs is just a theoretical speed, and it is actually much slower in real life, because it keeps on on having data loss.
Re:Will the wires catch on fire? (Score:5, Informative)
Oldschool HDMI was 4.2 Gbit. The newer 2.0 spec [wikipedia.org] is 18 GBit, which are almost certainly the most common now.
But you are correct (in spirit - we're talking Gbit, not GB) that the next gen, 2.1, claims to be 48Gbit, but a) they're not in wide enough use to test this argument (I don't know that any consumer gear has 2.1 yet), and b) you'd need to actually use that bandwidth (e.g. 4K/120, 8K), which again is not going to be common for some time.
It does look like HDMI 2.1 cables [amazon.com] are thick enough to have decent shielding.
Re: (Score:2)
Re:Will the wires catch on fire? (Score:5, Funny)
Nothing an $80 4ft cable with gold plated connectors couldnt fix!
Re: (Score:2)
yes, but what if your electrons aren't flowing in the right direction? how does monster get around that technical limitation/hurdle????
Re:Will the wires catch on fire? (Score:5, Funny)
Re: (Score:1)
Re: (Score:1)
I once pointed out to a manager that a 1,000-feet serial cable coiled up underneath the desk was unnecessary when a 20-feet cable could have connected the PC to the modem. He said it still work and that was that.
That is 20 times the maximum cable length of an RS232 cable! "Cable length is one of the most discussed items in RS232 world. The standard has a clear answer, the maximum cable length is 50 feet, or the cable length equal to a capacitance of 2500 pF." https://www.lammertbies.nl/com... [lammertbies.nl]
Re: (Score:1)
Re: (Score:2)
I've seen many out-of-spec cables work. But also, RS232 is hardly the only serial protocol out there; I've installed RS422 cables to connect equipment on opposite sides of a football stadium, all well within spec.
Re: (Score:1)
Re: (Score:1)
Re: (Score:1)
Re: (Score:1)
Re: (Score:1)
Really chris?
I'm not Chris.
What year was this?
Late 1980s.
Because RS-232 serial isn't supposed to go beyond 50 feet.
If you want the max speed at 19,200 baud. A 1,000 feet cable can run at 4,800 baud [db9-pinout.com]. For a 2,400 baud modem located 20 feet away, it worked just fine. That type of cable was for 500 feet or less runs for 9,600 baud between serial consoles and mainframe.
Re: (Score:2)
Re:Will the wires catch on fire? (Score:5, Informative)
Though a bit of an oversimplification, USB4 is basically just a rebranding of Thunderbolt 3. Thunderbolt 3 already does 40Gbps and has been out now for a few years. I have yet to hear reports of cables spontaneously erupting in flame or whatnot, and though USB 3.x and TB3 cables are stiffer than USB 2 cables, I don't think they're swaddled in insulation to a crazy degree. If you're curious how this will work, look back over the documentation for TB3.
Re: (Score:2)
Except Thunderbolt requires active cabling, and always has
Re:Will the wires catch on fire? (Score:5, Informative)
Unless USB4 cables will have a ton of insulation, to prevent the outside world from interfering with it. Or will it have more error checking thus this 40Gbs is just a theoretical speed, and it is actually much slower in real life, because it keeps on on having data loss.
You have to overcome miller capacitance in the cables, so the voltages are extremely-low.
What you do, you twist pairs of signal-carrying cables around each other, and you raise a signal cable by a few millivolts to signal. The signal pair will be e.g. 5mV apart. If you get EMI, then each cable will raise its voltage state equally, so you go from 0mV/5mV to 27mV/32mV. That's still 5mV, it's still signal, it's still clear.
Self-shielding.
Re: (Score:2)
Oversimplification, but yes this is one of several tricks that engineers uses to improve signal quality. Another commonly used one is ... shielding. In many cases you would use both. A good example of a cable that uses both is ... a USB 3.0 cable.
Re: (Score:2)
SATA and eSATA are the same interface, but not the same cable. SATA is inside a high-EMI envelope within a computer casing; eSATA is exposed to less electromagnetic interference.
SATA doesn't have shielding; eSATA requires shielding.
eSATA will work without shielding; so will USB 3.0. Both will also emit large amounts of EMI outside the shielded envelope, interfering with other electronic devices and violating FCC regulations.
USB 3.0 cables aren't shielded from outside electronics; outside electronics a
Re: (Score:2)
SATA and eSATA are the same interface, but not the same cable. SATA is inside a high-EMI envelope within a computer casing; eSATA is exposed to less electromagnetic interference.
SATA doesn't have shielding; eSATA requires shielding.
You description of the cables is right but the motivation behind them is quite different. Let's address them:
Firstly the EMI envelope within a PC is controlled and far lower than what you compare it to. Short lengths of very low current very low voltage signals at high frequencies radiate but do so poorly. On the flipside you have eSATA, a standard which will be routed directly next to unshielded LV power cables of multiple devices using a standard that allows double the length of cable, and will typically
Re: (Score:2)
Firstly the EMI envelope within a PC is controlled and far lower than what you compare it to. Short lengths of very low current very low voltage signals at high frequencies radiate but do so poorly.
The PC envelope is specifically designed to be an EMI shield due to EMI generated by the PC. It's an FCC compliance point.
And that is not remotely true. The data lines are twisted to prevent radiation and have been for a long time.
Twisting the data lines causes them to self-shield against near-end cross-talk (for round-trip pairs e.g. Ethernet, they'll have opposing magnetic fields which self-cancel), and also causes LVDS pairs to remain at the same base voltage when acting as antenna (these pairs don't self-shield against NEXT). It doesn't prevent them from radiating outward in an LVDS setup.
and in fact USB 3.0 all things being equal would be less likely to cause external interference than USB 2.0 based on signalling alone.
Oh really [intel.com]?
With the HDD connected, the noise floor in the 2.4 GHz band is raised by nearly 20 dB. This could impact wireless device sensitivity significantly.
With
Re: (Score:1)
Re:"...the mother of all wired connectivity option (Score:5, Interesting)
I don't think the "universal" is undeserved.
You used to have to have parallel and serial and PS/2 keyboard and mouse ports on your computer.
Now you can run almost anything - even direct hard drives, network adaptors, displays, telephones, modem, Wifi, 4G, etc. all of the same USB bus. To the point that machines can easily be supplied with nothing more than power and USB and still be fully functional. That's pretty "universal" to me.
The problem comes when people come up with competing standards - like Thunderbolt - which aren't part of the spec where your only option is to fold it into USB and basically have it be "Thunderbolt over USB". Fact is... it's universal enough that they can do that.
USB is pretty amazing. USB2 was just - to the amateur eye - faster versions of USB.
I can still plug in a mouse from the 1990's into a modern laptop and it "just works". It's only the oddball devices (which a universal specification allows - someone can easily make all kinds of nutty things that rely on OS-specific drivers, etc.) that don't and usually only because of issues unrelated to the USB transport itself.
USB is pretty damn good for what it is, and underappreciated nowadays.
Re: (Score:2)
USB bus
Universal Serial Bus bus?
Re: (Score:2)
Re: (Score:2)
Re: (Score:2)
Re:"...the mother of all wired connectivity option (Score:5, Insightful)
My main complaint with folding Thunderbolt into USB is the fact that it opens the door for asshole manufacturers like Apple to turn around and make laptops with a single fucking USB port, then expect consumers to go out and buy an ungodly expensive hub to unwrap everything.
In the beginning, there was DisplayPort. Using it with multiple displays required an expensive hub, but you could also use it with a cheap passive adapter cable to connect a single HDMI display. And it was good.
Then came Thunderbolt. Thunderbolt was multiplexed into DisplayPort. In theory, using the port for BOTH Thunderbolt AND DisplayPort required a UNGODLY expensive hub, but in reality, the only thing anyone cared about connecting via Thunderbolt was an external video card that had a DisplayPort or multiple HDMI ports of its own, so you could skip the expensive hub since you were still only connecting a single device to the computer itself using the computer's single DisplayPort port. And it was still good.
Then someone got the idea of multiplexing DisplayPort into USB. At first, it seemed like an OK idea... you could still use a cheap adapter cable to connect a single Thunderbolt eGPU to one of the ports, and use a $15 USB hub to connect things to the remaining USB port(s). After some nervous concern, it was still good.
Then Apple decided to Boldly Innovate, and sell laptops with a single USB port that needs a $500 hub if you want to use BOTH Thunderbolt (or DisplayPort) AND USB peripherals, and other manufacturers quickly followed just because they all blindly follow every stupid trend Apple comes up with. And it really, totally, fucking SUCKED.
Condensing everything -- PCIe, video, and USB -- into a single port that needs an expensive hub is OK when you're talking about a device like a phone that has extremely limited space for external expansion ports AND where using external peripherals is itself an extreme, rare edge case... but doing it with something like a LAPTOP where there's MORE than enough room for a half dozen ports, and would add only a few cents to the manufacturing cost, is just plan mean and user-hostile.
Yeah, combo hubs will probably be cheap SOMEDAY... but in the meantime, we're looking at 3-5 years of needing hubs that cost more than the peripherals connected to them. DisplayPort got away with needing an expensive hub, because most people didn't actually NEED that expensive hub to use it for the most common use case (connecting a single monitor). That's absolutely NOT the case with USB... especially if the manufacturer decides to pull an 'Apple' and ALSO use that single USB port for power delivery as well.
Re: (Score:2)
So buy a laptop with more ports.
I've got a laptop with 1 thunderbolt port (and two USB C), and damn, it's nice. I can charge from either side of the laptop, which is great for safety and comfort. I can plug two 4K monitors into one port with a pretty cheap dongle. It even works in Linux. And I can plug an external video card, but I've not tried yet.
It's extremely cool that I can expand a small, light, long battery life laptop into a configuration that's got all the comforts of a high end desktop.
Re: (Score:2)
Indeed. Fortunately unlike Apple and Microsoft other manufacturers choose to continue to compete on the desires of consumers and while that happens you'll continue to see multiple ports on laptops.
Re: (Score:2)
The biggest standardization effort was actually not on the hardware side. While USB has some nice features like hot-plugging and auto-negotiation it's basically readPacket() and writePacket(). The huge difference was that they started defining device classes like keyboards, mice, game pads, memory sticks, headphones/speakers etc. so you didn't need special drivers for each device. Some things took longer than other, like for example webcams took a while. You could watch the raw USB packets but the protocol
Re: (Score:2)
The problem comes when people come up with competing standards - like Thunderbolt - which aren't part of the spec where your only option is to fold it into USB and basically have it be "Thunderbolt over USB".
USB3 is a competing standard to USB1/2. It looks "universal" from a distance because the USB3 pins are hidden next to the old 1/2 pins. It's not even "USB3 over USB1/2", it's a bag on the side. The result is as universal as a lump of serial, parallel and PS2 connectors glued together -- sure, one of them will probably fit, but it's not really a solution to the multitude of different connectors if you just hide them all inside the same ground shell.
USB was supposed to replace "legacy ports" but now it has
Re: (Score:2)
> My solution? Call different technologies by different names. Make a different connector for different electrical protocols.
How do I benefit from this? I have a limited footprint for connectors in my laptop. Better to have electrical/pinout magic to make ports work with "slow" (but still pretty capable) devices, and multiplexer magic to allow multiple kinds of high speed signalling over the remaining pins.
These ports with multiple functions aren't too expensive, and I am usually better off with 4 "USB
Re: (Score:2)
Re: (Score:2)
Never mind that, what about IRQ attacks?
Re: (Score:2)
What about DMA attacks?
That's exactly the right question. Possibly they've got this worked out with TB authorization - supposedly the Windows from last April and boltd on linux do the right thing.
We've heard that Macintosh still allows stealing network traffic with DMA attacks, but maybe they can fix their IOMMU implementation. I do think Apple has the capability to fix it, but it's also possible that some of the early Thunderbolt machines don't get updates anymore.
I will be surprised if neither of the cu
Naming (Score:5, Funny)
So double the speed of 3.2... But which 3.2? Will it be equal to 3.2 2x2 or twice that?
Re: (Score:2)
It will be marketed as "Ludicrous Speed USB"
Re: (Score:2)
At this point, I find that to be the only acceptable move forward to be honest.
Re: (Score:2)
And the next one after that is going to be named "Plaid USB", obviously.
Re: (Score:1)
Did you mean 3.2 High speed or 3.2 Full speed?
Re: (Score:2)
Did you mean 3.2 High speed or 3.2 Full speed?
It's USB 3.2 Electric Bugaloo you insensitive clod
Re: (Score:2)
They should have said "fuck everything!" and gone straight to 5.
Re: (Score:2)
Re: (Score:2)
Latency? (Score:1)
Bandwidth is all very good but past a certain point, latency is of more interest. John Carmack's Tech Talk went into quite some detail [youtube.com]. So, wow us with the headline figure but it's not the whole story.
Re: (Score:2)
The latency for USB 3 is something like 30 microseconds. Some people are asking how anyone achieves that because they're only able to get as low as 60-70 microseconds.
That's like 250-550 signals per 60fps frame. Your input lag is 1/500th of an NTSC frame.
Re:What the fuck happened to "just works"? (Score:5, Funny)
USB 4 will once again ... (Score:2)
Re: (Score:3)
Re: (Score:2)
The cables and controllers sure aren't compatible, you need a converter from USB-x to USB-2 which typically includes a full controller on-chip. Even USB-3 to USB-2, the controllers aren't backwards compatible and thus older software that talks to OHCI/EHCI won't talk to xHCI or beyond.
Re: (Score:2)
???
> The cables and controllers sure aren't compatible, you need a converter from USB-x to USB-2 which typically includes a full controller on-chip.
USB3 type A has a few extra contacts, but you can plug USB1.0 type A cables into it and use USB1.0 devices, or anything inbetween...
> Even USB-3 to USB-2, the controllers aren't backwards compatible and thus older software that talks to OHCI/EHCI won't talk to xHCI or beyond.
Yah, you need an appropriate driver and OS that knows how to talk to xHCI, but act
Complexity is awesome! (Score:3)
One cord to rule them all!
What I really need to know is if it will support 3 phase 480V to run my HAAS CNC?
Network and storage over USB4 (Score:2)
However it is not quite being designed for true network and storage connectivity. The need is there but the buffers/latency might hurt it a bit.
I wonder if we'll see dedicated PCIE cards with dedicated USB4 chips so we can have nonblocking shorter-distance n
Re: (Score:2)
At these speed, USB4 stands to be way faster than most ethernet out there. 10gbit ethernet has been taking forever due to SPF power issues and manufacturing costs. USB4 with dedicated chip/channels would be twice the speed of full duplex ethernet 10gbit. However it is not quite being designed for true network and storage connectivity. The need is there but the buffers/latency might hurt it a bit. I wonder if we'll see dedicated PCIE cards with dedicated USB4 chips so we can have nonblocking shorter-distance network connectivity better than gigabit ethernet Or even USB-over-fiber next. Now that's a thought.
At these speed, USB4 stands to be way faster than most ethernet out there. 10gbit ethernet has been taking forever due to SPF power issues and manufacturing costs. USB4 with dedicated chip/channels would be twice the speed of full duplex ethernet 10gbit. However it is not quite being designed for true network and storage connectivity. The need is there but the buffers/latency might hurt it a bit. I wonder if we'll see dedicated PCIE cards with dedicated USB4 chips so we can have nonblocking shorter-distance network connectivity better than gigabit ethernet Or even USB-over-fiber next. Now that's a thought.
<sarcasm> Yeah, thanks to Tnunderbolt, one of those crap-ass ideas Apple came up with and never amounted to anything because they never became 'mainstream PC tech'. </sarcasm>
Presumably all those Apple haters out there will maintain full self-consistency by boycotting any hardware incorporating USB4 since it now integrates EEEEEVIL Apple tech!
Re: (Score:2)
Presumably all those Apple haters out there will maintain full self-consistency by boycotting any hardware incorporating USB4 since it now integrates EEEEEVIL Apple tech!
Hopefully they know Thunderbolt (Light Peak) is an Intel thing.
Re: (Score:2)
Looks like Thunderbolt was developed by Intel with collaboration from Apple. Not the other way around....
https://en.wikipedia.org/wiki/... [wikipedia.org]
the usb to e-net box will need an Chip and usb max (Score:2)
the usb to e-net box will need an Chip and 40 will max out the usb bus.
Also in servers needs to be cpu pci-e bus not stacked off of the PCH.
Re: (Score:2)
At these speed, USB4 stands to be way faster than most ethernet out there.
Until someone plugs a slow device into the USB bus. And your 40 Gbps network has to wait for those mouse data packets. This may be why vendors aren't in a rush to adopt USB for every application. Lots of pissed off customers because they don't understand that it's the old generation junk that they keep plugged in that slows down their fanc external hard drive.
Re: (Score:2)
Re: (Score:2)
I've been using 10Gbps Ethernet for a decade, and 40Gbps Ethernet (effectively four bonded lanes of 10Gbps) for half a decade. 10Gbps NICs are cheap now. You can run far longer distances than USB, and switches, routers, etc. are readily available. USB3 isn't going to replace Ethernet in the datacentre any time soon.
Be stuck with shit on board video? so no amd high (Score:2)
Be stuck with shit on board video? so no amd high end cpu's?? Need an video loop back cable?
Re: (Score:2)
but most AMD cpu's don't have any on board video. workstations are the same way.
Also you don't want USB bandwidth to eaten up by an HDMI display.
And Will cards pass more then 1 DP bus as there are limits to the number of displays
40 Gbps on a 40 Gbps cable? (Score:2)
too bad the usb c female connectors last ~ 1yr (Score:1)
I have had 2 phones with usb c now (nexus 6p and xaomi mi max 2), the usb c port has lasted about 1 year before getting loose and having to be propped in a certain way to connect, It is not a good port in that respect. I wish they would make a better solution.
Re: (Score:2)
And yet here we are with two decades old hardware that still have perfectly functional USB type A ports.
Re: (Score:2)
Re: (Score:2)
as it became obvious that it's a bad idea to allow someone to plug something in that can be both a storage device and a keyboard (it doesn't take much imagination to see how that can be a security problem.)
We used to have key-capture devices like that to slip between an AT keyboard and the port.
There's a reason I described voting machine standards in which no physical electrical port may be accessible between polling begin and end of polling day after generating proof of ballot set. We're going to have to go into glorious battle to force vendors to accept these standards, but I'm ready for that.
Just great. (Score:3)
Re: (Score:2)
Correct. And it will have USB-C connectors so it will look identical to a USB 3.2 cable next to it in your parts drawer. And since the 3.2 cable will probably work, albeit at a lower speed, good luck ever figuring out which is which. Yay standards?
Re: (Score:2)
Yes yay standards. I can't find my USB 3.2 cable so I can just use my thunderbolt cable. That's what standards allow.
If you have a problem with the confusion why would you suffer by leaving obsolete crap lying around. You can just thunderbolt the world :)
Re: (Score:2)
Re: (Score:2)
So now I need to buy another fucking cable.
No you don't. The first device you buy which needs this will come with the cable.
Seriously who has ever bought a USB cable. Weird.
How many flavours? (Score:2)
To empower consumers, USB4 will come in a variety of new flavours: /w ABS /w cheese
-USB4
-USB4 gen 1.5
-USB gen pi
-USB4 2x2
-USB4 4x4
-USB4 4x4
-USB4
-USB4 ultra graphics pro turbo
-USB4 with kung-fu action grip
Of course, you have no idea what particular flavour your cable will be and if it will be compatible with the devices you are connecting, but hey, it's progress!
Re: (Score:2)
I'm hoping they move entirely to Type-C and establish HDMI-over-type-C, audio accessory mode, Ethernet alternative mode, and high-speed charging in the base standard.
Re: (Score:2)
usb4-swift = usb 1.0
usb4-quick = usb 1.1
usb4-brisk = usb 2.0
usb4-express = usb 3.0
usb4-hasty = 3.1
usb-4-dashing = 3.2
Re: (Score:2)
Where's my high speed LAN? (Score:2)
Great. Now my keyboard can do 40 Gbps, but my LAN is still stuck at 1 Gbps. Why can't this inexpensive technology be used to give me a high speed LAN? 10GB Ethernet still costs thousands of dollars.