Gigabits per second should be expressed as Gb/s, not GB/s. I have seen a lot of instances of improper units being used in tech media, and it would be better for everyone if units were expressed correctly to improve the clarity of content.
I dont even think the majority of people will know the difference or ever care. I mean the entire gibibyte thing is pretty recent, and only came about after the FSF advocated for its usage, its pretty much redundant and tons of people use the terms interchangeably because for the most part it doesnt matter. on a little further research, there seems to be no real clear consensus on which stands for which.
To be even more accurate we shouldn't use Giga since it's the term for 10-base numbers. The correct unit is Gib/s pronounced Gibibit...at least if we mean the binary values.
I mean psvr2 will do it and it will do 6dof tracking and eye tracking via that usb ......and all in ones can go wireless or wired via usb ...it might not be optimal but is more than great and it will be way better in the future
Or how about USB 80Gbps Pro and it supports everything, and then there is also USB 80Gbps Mobile which is only required to support the speed and data requirements. PCIe tunneling and extra stuff can be omitted for the Mobile version
In all seriousness though, if you have the correct cables and motherboard/pc, it's incredible the age we're entering into. External gpus and drives are becoming faster than internal.
I think it is intentional, so that manufacturers of laptops or pcs can advertise that they have USB 3.2 and the customer will buy not knowing there are 5 variants and that one is the slowest one
Even if the names weren't confusing at all, half of the specs are optional on each iteration. Which means you might as well not have a standard, since you'll still need to check the listing of each individual spec of each individual port. God awful stuff
Thank you for mocking the hell out of that “USB 4 version 2” name, it’s ridiculous and must never materialize in consumer products. Just call it USB 5 for humanity’s sake
Tech quickie are misleading you. It's not "USB 4". Rather a single word "USB4". That's why they were able to add a version number 2 to the word USB4. this is a non-final name and the likes of LTT and other influencers are being angry before the final specification is revealed.
@@YounesLayachi the only thing customers should ever see now is "USB4 80Gbps" or similar. I wish LMG would do their homework on this instead of peddling this FUD. This comments section is full of brainless comments complaining about an issue that doesn't exist.
@@roanv sure, thatd be better version numbering, but USB4 80Gbps is better for consumers, since both the 40Gbps and 80Gbps can use the same cable. basicly, id say that as long as a new cable is not requiered, the version number should stay the same, just whit the speed being upped, and if the speed increases isint huge but the cable still chanched, then it should be a 4.1, similar to 3.1 and 3.2
@@mr.battlecats5512 naming is the same for devices/ports and cables. As I see it there are only two things that may prevent manufacturers from correctly labeling cables: aesthetics and cost.
I remember the time when somebody added USB 4.0 version 2 support patch to Linux Kernel, and the maintainer (Greg Kroah-Hartman) be like "wtf 4.0 version 2 meant for? Please use common sense for versioning" aand he realized it's part of the standard lol
PAM (Pulse Amplitude Modulation) is a catch-all term meaning you change the amplitude of the signal to carry data. PAM4 (PAM with 4 amplitude levels) is the specific scheme described in this video.
how high can PAM go without causing problems? as in, can we have PAM16 or PAM64 no problem? Will my future 240 W USB-C charger also possibly have 80 GB/s signal so that I can power a low power desktop computer while also connecting it to a TV and having it take power from its wall?
@@AR15ORIGINAL It depends on many factors, the main one being related to noise. You need enough separation between amplitude levels to transfer data reliably, or else noise will induce errors. Errors are expected and can be mitigated by parity bits or better cable construction, but this increases cost.
@@AR15ORIGINAL I missed this message! As gustavrsh mentioned, it becomes exceedingly hard to reach higher levels of PAM. It's already hard enough to hit 40Gbps over copper which is why copper TB4/USB4 cables tend to be very limited in length. There is too much resistance over the copper that the switching frequency can't be efficiently increased. With PAM, this becomes a bigger issue because when you used to have, say, a 5V potential difference from a digital 1 and digital 0, now you have four levels for 00, 01, 10, and 11. That means you can only get a potential difference of 1.25V between each level. Your margin for error is significantly diminished, further reducing the potential length a copper cable can reach before there is too much signal integrity lost. There are ways around this, either build more and more advanced copper cables that have shielding and such (like what ethernet continues to do being able to reach 10Gbps over a 100m cable with Cat6A), or adopt optical cables (which is where datacenter networking has gone). There are practical limits to what copper can do and many people believe silicon photonics are the future, where even things like sending data from your CPU to your GPU or RAM will be done over optical cable because it gets harder and harder to promise double the bandwidth every 2-3 years in the same form factor. As gustavrsh also mentioned, QAM is a good technology that can increase bandwidth far beyond the capabilities of PAM. This is done by assigning each value to a phase and amplitude. So for QAM16, you might have 0000 assigned to amplitude 25%, phase 45°. By placing all the values on a grid like this, you can squeeze a lot more values in without needing to increase your error rate so much. However, there is a big catch. Since you're sending phase-shifted signals, this is not used in electrical signaling at all afaik. This is because it is nearly impossible to send data at high bitrates while also being able to pull the voltage to the desired value quickly enough. When placing two phase-shifted values next to each other, you often get discontinuous signals, where one signal ends at, say 75%, while the other starts at 25%. This can be handled by wireless systems easily since they are just transmitting data over the air with no voltage to change over resistive wire. But when running those extremely high frequencies, you'll have already overlapped with the next bit by the time you get done pulling your voltage to the correct value and sending the appropriate data.
CORRECTION: What is described here is PAM4 as in there are 4 levels that the signal can be at. The older way of doing things can be considered PAM2, but is usually called NRZ.
well what prevents them from doing an even bigger increase to have it at 8 ( 000 , 001 , all the way to 111) or 16 levels (..), or is it not that easy ?
@@Ibrahim1999 I believe it would be related to the accuracy of signalling, because it increases the complexity of the signal by having multiple voltage levels instead of just on/off or positive/negative. More levels means more accuracy required, increasing cost of the hardware required. It introduces problems with signal integrity, because voltage drop and the voltages affecting one another can change the signal. This is why PCIe Gen 4 and especially Gen 5 motherboards and such are more expensive, have more PCB layers, and so on
USB 4.20 sounds great on the paper, if it really is that good then hopefully it will be widely available in consumer electronics in the next year or two
Version 2 is just the updated USB 4 specifications, that's not going to be used for marketing or any product labels. The USB 4 speeds will be labeled as 20 Gbps, 40 Gbps and 80 Gbps
I'm simply amazed that multiple people can spend hours and hours of their limited time on this planet to make videos like this without doing basic homework first.
I'm still confused why they never change the physical connector to begin with I thought that was the whole point of introducing USB C but nope most computers still have USB A I don't think I've ever seen a single desktop computer with all USB C.
@@CalculatedRiskAK INTEL,That's the Problem ! EVERYTIME I see a GREAT 4K monitor now,with ALL the Specs & Ports I want,it "DOESN'T" have the the USB-C connector on it. ESPECIALLY when the NVIDIA name is tagged to it ! DAMIT !! AND if they do,they cost you that Kidney + a Left NUT ! If it does have the Connector,they shorted it somewhere,ALWAYS !😎
I like how the USBIF came up with a far more archaic and nonsensical naming scheme than Daru in Stein's Gate, especially considering that Daru's was _designed_ to be archaic and nonsensical
the 120 Down 40 Up will be implemented in USB 4 version 2.1 alpha. They're literally becoming the Street Fighter series in naming conventions. Street Fighter 3 Alpha Turbo Championship Edition.
Love the content. but! , *pushed up glasses* , Actually this is referred to as PAM4, not just PAM. PAM3, Pam 5 and other types of PAM are all available. Some guy even made a PAM16. PAM is just Pulse Amplitude Modulation. Source: I am a hardware verification engineer for, among other things, high speed (20 Gbs) memory related asics.
I wish usd plugs had stronger clips that kept the plugs more snug in the port. Or they made the port connections stronger so they don't wear out so fast.
thing is, the harder it clips in, the faster itll ware out bc itll have more pressure pushing those clips every time you insert it or take it out. its a balancing act. also you want the wire to ware out faster than the port, bc the wire is easier anc cheaper to replace then the thing its attached into.
It feels like we can fix this naming mess. What if engineering degree programs include some rules for nomenclature or something? maybe that could even just be separate IEEE standard?
1:33 proof you can wear sandles anywhere.... not that you should. I do like that they are looking into allowing USB to be nonsymetrical given most USB devices don't need symmetric RX and TX.
So eventually my PC case will have a USB5 hub that all my SSDs and GPU plug into and the power supply will just be a connection for that hub and the MoBo connectors.
They could've just name it USBx Catx. If they found way to increase the bandwidth but still not enough for next gen, they can just add the Cat number like ethernet does.
although transfer speed will be doubled, the new signaling scheme also means that there will be a performance overhead needed for re-decoding which is not a small depending on the use case specially for high bandwidth applications such as HD video rendering... but a good progress nonetheless since USB is severely bottle-necked compared to CPU and GPU performances... there is also another option to add that function to the USB BUS since it does absolutely almost nothing but bottle-necking high performance usb devices right now... a separate processor for the bus "MIGHT" mean that they could actually go more ham on the USB speeds but further increasing/complicating the signal scheme and let the new dedicated processor do the decoding... but this also means added costs for computers which are horrendously expensive right now
umm, obviously its going to be "decoded" at the usb bus lol, wtf man what would be the point of this when you can just compress data anyway for transport
@@Sup3rman1c oh god. are you sure you're talking about the BUS? the bus has 3 functions... 1. read/write which is the controller part, 2. address which decide where the data is sent, 3. data which are split in lanes which also decides speed and how much of it are allocated to each USB device but the data itself including compression and coding are still handled by the processor... metaphor, the processor is your brain, the bus is closer in function to your spine and nerves... usb 4v2 coding in the video uses a 4 signalling system which needs to be reconverted back to 1 and 0, in theory, you can still over complicate that with more than 4 signals (8, 16, 32... so on, yes it also has to double since theyre pretty much derivatives of 1 and 0 scheme but wont happen anytime soon without more sensitive hardware) which would need a separate processor if you don't want an overhead processing requirement from your CPU which in theory may or may not be needed since CPU's today are really fast... this is what i was talking about. its theory moving forward on what could happen. food for thought, processors can be anything other than the CPU and GPU. its any piece of circuitry that can convert signal/inputs into instruction sets the the computer can use... The CPU in itself is made to process a very diverse set of functions but keep in mind that there are processors that are made to do only 1 thing and do that one thing really well. another food for thought, you do know data compression is also still the function of the CPU...
Feedback - While Presenting Riley pointed out correctly about speed being 80Gb/s and 40Gb/s but the graphics shown were of 80GB/s and 40GB/s. Bits and Bytes matters
A single USB no-compression VR headset would actually be a pretty great use case for the high bandwidth. The bulky proprietary cables is not only a factor discouraging people from trying PC VR, but also a major point of failure and potential extra cost down the line. A single cable would be a much more attractive option. You can even buy several third party cables and attach one to every machine you want to use the headset with...
@Erbkaiser it's called "USB4" without the space *specifically so that can't happen*. Version numbers like 4.0 can't be trademarked, but "USB4" can, so if a device is mislabeled they can take legal action against that.
Once they get to a point where you can plug in portable USB graphics cards, it will open so many opportunities, especially since laptops aren't upgradable for the most part.
@@Big_Tex Glad someone in this space has some humor 😂 it seems like the goal of every company is to see if they can cause some confusion with their naming schemes
The USB naming issues have long been obnoxious, but with USB4, they are trying to establish a better convention and this video could have educated people instead of adding to the confusion. Version 1.0 of the USB4 specification defines a series of ports, including USB4 20Gbps and USB4 40Gbps. Version 2.0 of the USB4 specification adds an additional port: USB4 80Gbps. There are internal names based on how each mode works, such as USB4 2x2 and USB4 3x2 for 20 and 40 Gbps respectively. The spec also defines some other variations, as well as carries forward the definitions of the 3.2 spec. On the whole, we would all benefit from referring to the ports by their speed going forward and should stop referring to them by the version number of the specification.
Another way to easily increase the speed is using PCIe Gen 4 (probably they are using this), the way Thunderbolt 3 and USB4 archive 40gpbs is using 4 lanes of PCIe 3.0 giving them around 32gbps (which by the way is the actual amount of speed you could use to transfer data in like an eGPU or everything that uses the PCIe lanes, In fact I'm not sure you can use those 40gbps for storage) so using 4 lanes of PCIe 4.0 would leave them almost all the way there, in fact should be easy to upgrade the USB or Thunderbolt standard to over 100gbps using PCIe 5.0 lanes, I assume they didn't go offering that due to power consumption and cost (reasons why laptops are still beginning to adopt PCIe 4.0, suspiciously good timing to implement "USB 4.1" speeds) probably also the thing that prevents them to bump up the speed following PCIe standard is getting that speed over some distance with a USB-C cable
1 cable for everything sounds awesome. I'm not sure why we still need Ethernet, display port, HDMI, 3.5mm, and coax? Ethernet and coax being the strangest ones.
Honestly the the thing I'm looking forward to is the idea that a laptop can be your primary gaming device and when you get home have 90+ percent of the GPU, keyboard, and mouse ready to go on a single dock cable.
OMG they can simply solve both problems by just simply having different levels meant for different things: 4.1 Minimum requirements aimed for phones 4.2 Higher requirements aimed for tablets or notebooks 4.3 Mid level requirements aimed for Laptops and consoles 4.4 High level requirements meant for powerful laptops or desktops 4.5 Max requirements aimed for top end laptops or powerful desktops Then when someone advertises 4.4 features but only have 4.2 features, you will actually this time be able to tell the difference. I didn't put much thought into the specifics but you get the idea
And besides the one cable ideal, wouldn't it be nice if we could backup an internal M.2 at 10 GB/s? This is why more than one USB 4 ports eventually are desirable as well.
So I've been checking online for a USB C splitter that can connect a 3.5mm Jack and a USB C. And from what I've found they have limited compatibility with some phones or has some variation (18w, 30w, 60w charge, no data transfer etc) so why is this? I thought buying a USB C at this point is a no brainer, but now you have to sift through those variations just so you can buy the right best working one.
I am thinking of a future where pcs just have like 50 type c ports on the back and you can connect any device into any port and it'll just detect what it needs to do Except for the dedicated GPU display outputs, they could be type c but will need to plug displays into them rather than "any" of the ports
You know, if they _had_ used the cooking spray, they could've gotten even more speed by keeping bits from sticking to the inside of the wires.
😂😂😂
i didn't get the joke
Dad?
@@realfoggy milk?
That only works for fibre optical cables.
Gigabits per second should be expressed as Gb/s, not GB/s. I have seen a lot of instances of improper units being used in tech media, and it would be better for everyone if units were expressed correctly to improve the clarity of content.
I dont even think the majority of people will know the difference or ever care. I mean the entire gibibyte thing is pretty recent, and only came about after the FSF advocated for its usage, its pretty much redundant and tons of people use the terms interchangeably because for the most part it doesnt matter.
on a little further research, there seems to be no real clear consensus on which stands for which.
80GB/s is around 640gbps so it's misinformation.
@@azzzertyy I agree, but I think if everyone that created media cared, people would start caring and we could start communicating more effectively
@@jcnicholls2131 80GB/s is exactly 640gb/s
To be even more accurate we shouldn't use Giga since it's the term for 10-base numbers. The correct unit is Gib/s pronounced Gibibit...at least if we mean the binary values.
I didn't even know there was a usb 4.0
It's not USB 4.0, it's USB4
@@jankeller1635 probably not going to be called that in a year or so, just wait...
Yeah lol same
same 😂😂😂
same
I'm sad that VR headsets didn't go for a single USB C port straight into the GPU.
you can do that with quest 2 and DisplayPort to usbc adapter
My 2080ti has that
VirtualLink I thinks it's called.
I mean psvr2 will do it and it will do 6dof tracking and eye tracking via that usb ......and all in ones can go wireless or wired via usb ...it might not be optimal but is more than great and it will be way better in the future
@@paineretlaw3344 It "exists", but Valve etc ditched it even though it was announced as a feature.
@@CanadianBakin42O you should go back to the English teacher
The full name should just be USB 4.1 gen2x2 v3 asymmetrical 40/120gbps type-c port with display port 2.0 and thunderbolt 3 support
Or how about USB 80Gbps Pro and it supports everything, and then there is also USB 80Gbps Mobile which is only required to support the speed and data requirements. PCIe tunneling and extra stuff can be omitted for the Mobile version
In all seriousness though, if you have the correct cables and motherboard/pc, it's incredible the age we're entering into. External gpus and drives are becoming faster than internal.
1:02 Riley's superpower of instantaneously changing shirt colors lol
Yep - made my inner Monk scream in agony 😖
Is it a private joke regarding the name changing of usb 3 and stuff ?
So he was using shirt 4 v2 gen3x2?
They *have* to know that the names are stupid and overly confusing, right? So at this point we can agree it's intentional
I think it is intentional, so that manufacturers of laptops or pcs can advertise that they have USB 3.2 and the customer will buy not knowing there are 5 variants and that one is the slowest one
Only losers start counting at 0 or 1, the cool kids start at 2.
Probably not, they get paid the big bucks because they're engineers, not smart.
Even if the names weren't confusing at all, half of the specs are optional on each iteration. Which means you might as well not have a standard, since you'll still need to check the listing of each individual spec of each individual port.
God awful stuff
@@paulelderson934 But it uses a (Colour port*). Takes 50% of the guesswork out of it. Or does it? Pay more to find out!
Every single time Riley grimaces due to naming convention:
1:59
2:32
2:54
4:28
4:55
It's not naming convention, it's naming chaos
Doing God's work right there
😅 1:43👍🏼
Someone make a short out of this.
His expressions just carried this video tbh
Thank you for mocking the hell out of that “USB 4 version 2” name, it’s ridiculous and must never materialize in consumer products. Just call it USB 5 for humanity’s sake
Tech quickie are misleading you.
It's not "USB 4". Rather a single word "USB4".
That's why they were able to add a version number 2 to the word USB4.
this is a non-final name and the likes of LTT and other influencers are being angry before the final specification is revealed.
@@YounesLayachi the only thing customers should ever see now is "USB4 80Gbps" or similar. I wish LMG would do their homework on this instead of peddling this FUD. This comments section is full of brainless comments complaining about an issue that doesn't exist.
@@jfolz or they could just be sensible and name it usb 4.1?
@@roanv sure, thatd be better version numbering, but USB4 80Gbps is better for consumers, since both the 40Gbps and 80Gbps can use the same cable. basicly, id say that as long as a new cable is not requiered, the version number should stay the same, just whit the speed being upped, and if the speed increases isint huge but the cable still chanched, then it should be a 4.1, similar to 3.1 and 3.2
@@mr.battlecats5512 naming is the same for devices/ports and cables. As I see it there are only two things that may prevent manufacturers from correctly labeling cables: aesthetics and cost.
I remember the time when somebody added USB 4.0 version 2 support patch to Linux Kernel, and the maintainer (Greg Kroah-Hartman) be like "wtf 4.0 version 2 meant for? Please use common sense for versioning" aand he realized it's part of the standard lol
Well, he's not wrong. They need to use common sense with versioning.
1:02 - 1:28 It's fun seeing -alternate reality Rileys- cuts put in on a later day when new information comes up.
Some time in the future.... USB 4.4.2 Gen 3.5 SuperSpeed Deluxe Pro
Don't You Put that Evil on Me, Ricky Bobby!
uhhh you missed the "Ultimate" not to mention the "Max"... pfff
Which version are you referring to exactly?
...ultra max extreme
Also don't forget about USB 5.4 gen 4x3 version 2 revision 3 superspeed max pro ultra plus deluxe (69Tb/s)
PAM (Pulse Amplitude Modulation) is a catch-all term meaning you change the amplitude of the signal to carry data.
PAM4 (PAM with 4 amplitude levels) is the specific scheme described in this video.
how high can PAM go without causing problems?
as in, can we have PAM16 or PAM64 no problem?
Will my future 240 W USB-C charger also possibly have 80 GB/s signal so that I can power a low power desktop computer while also connecting it to a TV and having it take power from its wall?
@@AR15ORIGINAL It depends on many factors, the main one being related to noise. You need enough separation between amplitude levels to transfer data reliably, or else noise will induce errors. Errors are expected and can be mitigated by parity bits or better cable construction, but this increases cost.
@@AR15ORIGINAL also if you want to go above 16 levels you may as well use QAM (WiFi 5 is mainly 256-QAM, WiFi 6 is 1024-QAM)
@@AR15ORIGINAL I missed this message! As gustavrsh mentioned, it becomes exceedingly hard to reach higher levels of PAM. It's already hard enough to hit 40Gbps over copper which is why copper TB4/USB4 cables tend to be very limited in length. There is too much resistance over the copper that the switching frequency can't be efficiently increased. With PAM, this becomes a bigger issue because when you used to have, say, a 5V potential difference from a digital 1 and digital 0, now you have four levels for 00, 01, 10, and 11. That means you can only get a potential difference of 1.25V between each level. Your margin for error is significantly diminished, further reducing the potential length a copper cable can reach before there is too much signal integrity lost. There are ways around this, either build more and more advanced copper cables that have shielding and such (like what ethernet continues to do being able to reach 10Gbps over a 100m cable with Cat6A), or adopt optical cables (which is where datacenter networking has gone). There are practical limits to what copper can do and many people believe silicon photonics are the future, where even things like sending data from your CPU to your GPU or RAM will be done over optical cable because it gets harder and harder to promise double the bandwidth every 2-3 years in the same form factor.
As gustavrsh also mentioned, QAM is a good technology that can increase bandwidth far beyond the capabilities of PAM. This is done by assigning each value to a phase and amplitude. So for QAM16, you might have 0000 assigned to amplitude 25%, phase 45°. By placing all the values on a grid like this, you can squeeze a lot more values in without needing to increase your error rate so much. However, there is a big catch. Since you're sending phase-shifted signals, this is not used in electrical signaling at all afaik. This is because it is nearly impossible to send data at high bitrates while also being able to pull the voltage to the desired value quickly enough. When placing two phase-shifted values next to each other, you often get discontinuous signals, where one signal ends at, say 75%, while the other starts at 25%. This can be handled by wireless systems easily since they are just transmitting data over the air with no voltage to change over resistive wire. But when running those extremely high frequencies, you'll have already overlapped with the next bit by the time you get done pulling your voltage to the correct value and sending the appropriate data.
@@gustavrshso we're back to analog?
I FUCKING KNEW IT! I KNEW ANALOG WILL BE BACK!
Science: "The simplest, most proximity to the truth"
USB-IF: "Most irritanting USB names, more USB is (un)famous"
0:28 He said 80 gigabits but on screen it showed 80 gigabytes
i saw that too and wanted to comment on that; that way i found your comment
CORRECTION: What is described here is PAM4 as in there are 4 levels that the signal can be at. The older way of doing things can be considered PAM2, but is usually called NRZ.
well what prevents them from doing an even bigger increase to have it at 8 ( 000 , 001 , all the way to 111) or 16 levels (..), or is it not that easy ?
@@Ibrahim1999 noise. It’s already hard to tell 1 from 0 in some bad HDMI cables, and every level above that makes it worse and worse.
@@Ibrahim1999 I believe it would be related to the accuracy of signalling, because it increases the complexity of the signal by having multiple voltage levels instead of just on/off or positive/negative. More levels means more accuracy required, increasing cost of the hardware required. It introduces problems with signal integrity, because voltage drop and the voltages affecting one another can change the signal. This is why PCIe Gen 4 and especially Gen 5 motherboards and such are more expensive, have more PCB layers, and so on
@@JerzyMuller TBF, PAM is kinda rudimentary. QAM64 was already a thing when I studied data communications in 1999.
👍🏼1:43😎✌🏼
USB 4.20 sounds great on the paper, if it really is that good then hopefully it will be widely available in consumer electronics in the next year or two
Usb 420 😂😂😂😂
Best comment here hahaha
@@topg2820 Don't shoot the messenger. I am only using the current naming scheme 😉
Version 2 is just the updated USB 4 specifications, that's not going to be used for marketing or any product labels. The USB 4 speeds will be labeled as 20 Gbps, 40 Gbps and 80 Gbps
I'm simply amazed that multiple people can spend hours and hours of their limited time on this planet to make videos like this without doing basic homework first.
Yep, even USB 3 have the same labels.
Nothing confusing about USB 4 vII. Just remember it’s the one after USB 4 Gen3x2
Teamgroup's original giveaway end date: July 20, 2022
Revised end date: July 31, 2022
Date this promotion airs: September 20, 2022
Moment of silent thanks for the straight forward Bluetooth naming scheme.
0:26 b is for bits B is for Bytes, know the difference, it might save your quickie.
Honestly as long as they don't change the physical connection I'm happy
I'm still confused why they never change the physical connector to begin with I thought that was the whole point of introducing USB C but nope most computers still have USB A I don't think I've ever seen a single desktop computer with all USB C.
@@bland9876 Well, USB3.2 Gen2x2 and above are USB C only, so it's a matter of time before the old USB A become obsolete.
2:39 maybe this will be included in the upcoming USB 4 v2 part 2
😂
So it's 80 gigabit not gigabyte right? First graphic showed gigabyte, but he said gigabit which matches what the display port shows later
He mentioned gigabyte later so infographic is definitely wrong. I was confused too.
You're saying 80 Gigabits per second while showing on screen 80GB/s. The capital B indicates GigaBYTES, whereas lowercase b indicates GigaBITS
100%. Tech channel should be getting this right.
We’re now waiting for the next concept of merging display cables into the USB fold. One cable to rule them all…
I mean, Thunderbolt can already do this. It's just very expensive for some reason. Not sure why
@@GT-tj1qg Thunderbolt is expensive because Intel licensing.
@@CalculatedRiskAK INTEL,That's the Problem ! EVERYTIME I see a GREAT 4K monitor now,with ALL the Specs & Ports I want,it "DOESN'T" have the the USB-C connector on it. ESPECIALLY when the NVIDIA name is tagged to it ! DAMIT !! AND if they do,they cost you that Kidney + a Left NUT ! If it does have the Connector,they shorted it somewhere,ALWAYS !😎
I'd be happy if my usb 3 worked correctly
USB rants are some of the best content on LTT channels along with paint rants and Nvidia rants 😆
I like how the USBIF came up with a far more archaic and nonsensical naming scheme than Daru in Stein's Gate, especially considering that Daru's was _designed_ to be archaic and nonsensical
El Psy Kongroo!
The next one is USB 4.0+ and not USB 5.
Also the U stands for Ultimate now.
I am mad scientist! Sonuvabeech!!!
Every time it makes ding the world of USB has changed.
@@Moody-Boy You had me laughing so hard for a minute. Thank you, it made my day.
Hopefully the marketing name is still based on the Gbps
USB4 20Gbps, USB4 40Gbps, and USB4 80Gbps
the 120 Down 40 Up will be implemented in USB 4 version 2.1 alpha. They're literally becoming the Street Fighter series in naming conventions. Street Fighter 3 Alpha Turbo Championship Edition.
Riley’s sarcasm speaks for all of us
0:30 tell whoever made that prompts that GB means GigaByte. It should read Gb with lowercase b so it means Gigabit
Love the content. but! , *pushed up glasses* , Actually this is referred to as PAM4, not just PAM. PAM3, Pam 5 and other types of PAM are all available. Some guy even made a PAM16. PAM is just Pulse Amplitude Modulation. Source: I am a hardware verification engineer for, among other things, high speed (20 Gbs) memory related asics.
3:28 I would like to know what constitutes a non-geometric silhouette.
I wish usd plugs had stronger clips that kept the plugs more snug in the port. Or they made the port connections stronger so they don't wear out so fast.
thing is, the harder it clips in, the faster itll ware out bc itll have more pressure pushing those clips every time you insert it or take it out. its a balancing act. also you want the wire to ware out faster than the port, bc the wire is easier anc cheaper to replace then the thing its attached into.
Can't wait to see USB 5 Gen4.5x3.3 v2.2+ comes in my next laptop. Think it will be powerful.
It feels like we can fix this naming mess. What if engineering degree programs include some rules for nomenclature or something? maybe that could even just be separate IEEE standard?
I like the visible pain of Riley when saying USB version 2.0.
1:33 proof you can wear sandles anywhere.... not that you should.
I do like that they are looking into allowing USB to be nonsymetrical given most USB devices don't need symmetric RX and TX.
Maybe we’ll get asymmetrical data transfer in USB 4 v2 release 5, third cohort, second gen 1.3H
No that'll be in USB 4.3 gen 3x2 version 3 revision 2.1 second release
Now we can use USB 4 flash drives as boot drives now!
You know if everyone called it USB 4.2 And ignored their naming convention. But stuck strictly to that. They would have to change it.
@1:30 Gb/s not GB/s. Use the correct symbols for bits and bytes. No quality control?
1:04 _Riley 4.0 Version 2 blue shirt_ 🤨
Now THAT is a version name! 😂
So eventually my PC case will have a USB5 hub that all my SSDs and GPU plug into and the power supply will just be a connection for that hub and the MoBo connectors.
Correction. It will have a USB 5.3 gen 4x3 version 3 revision 2.1. Jokes aside yes in the future the Universal part of the USB can finally be achieved
The great thing about standards is that there's so many to choose from.
Legendary performance mate, made the video 100 times better.
This shade is fantastic. Easily one of Riley's finest performances
They could've just name it USBx Catx. If they found way to increase the bandwidth but still not enough for next gen, they can just add the Cat number like ethernet does.
Great video, i was impressed with Riley's shirt changing speed at that one segment he wore light blue.
although transfer speed will be doubled, the new signaling scheme also means that there will be a performance overhead needed for re-decoding which is not a small depending on the use case specially for high bandwidth applications such as HD video rendering... but a good progress nonetheless since USB is severely bottle-necked compared to CPU and GPU performances... there is also another option to add that function to the USB BUS since it does absolutely almost nothing but bottle-necking high performance usb devices right now... a separate processor for the bus "MIGHT" mean that they could actually go more ham on the USB speeds but further increasing/complicating the signal scheme and let the new dedicated processor do the decoding... but this also means added costs for computers which are horrendously expensive right now
umm, obviously its going to be "decoded" at the usb bus lol, wtf man what would be the point of this when you can just compress data anyway for transport
@@Sup3rman1c oh god. are you sure you're talking about the BUS? the bus has 3 functions... 1. read/write which is the controller part, 2. address which decide where the data is sent, 3. data which are split in lanes which also decides speed and how much of it are allocated to each USB device but the data itself including compression and coding are still handled by the processor... metaphor, the processor is your brain, the bus is closer in function to your spine and nerves... usb 4v2 coding in the video uses a 4 signalling system which needs to be reconverted back to 1 and 0, in theory, you can still over complicate that with more than 4 signals (8, 16, 32... so on, yes it also has to double since theyre pretty much derivatives of 1 and 0 scheme but wont happen anytime soon without more sensitive hardware) which would need a separate processor if you don't want an overhead processing requirement from your CPU which in theory may or may not be needed since CPU's today are really fast... this is what i was talking about. its theory moving forward on what could happen.
food for thought, processors can be anything other than the CPU and GPU. its any piece of circuitry that can convert signal/inputs into instruction sets the the computer can use... The CPU in itself is made to process a very diverse set of functions but keep in mind that there are processors that are made to do only 1 thing and do that one thing really well.
another food for thought, you do know data compression is also still the function of the CPU...
Feedback - While Presenting Riley pointed out correctly about speed being 80Gb/s and 40Gb/s but the graphics shown were of 80GB/s and 40GB/s. Bits and Bytes matters
Riley's expression will never cease to make me crack up
Thanks for linking to a giveaway that seems to have ended in July
Could we possibly start to see high speed dual 40gbit/s networking via USB ?? Near 100Gbit/s networking on a laptop would be pretty cool IMO.
That's really a great use of that much of bandwidth.
I'm just glad it's at last Type-C only for USB4 so version 2.0 adding double bandwidth is a neat extra.
So cool. Can't wait for the day that everything is done over usbc.
0:27 what? 80 gigabits per second as he said? or 80 GB/s as it's written?
Wake me up when they have cables that can do 80gbps and 240w
What usecase do you have for that?
@@lioneckenbach552 power their SFF off a USB port ;_;
A single USB no-compression VR headset would actually be a pretty great use case for the high bandwidth. The bulky proprietary cables is not only a factor discouraging people from trying PC VR, but also a major point of failure and potential extra cost down the line. A single cable would be a much more attractive option. You can even buy several third party cables and attach one to every machine you want to use the headset with...
How did we get from "USB 2.0" to "USB4 2.0 Gen3x2".
Which normal person call tell the differences between the USB specifications.
The first gen is USB4, the second gen has a space after, so its USB 4 v2
@Erbkaiser it's called "USB4" without the space *specifically so that can't happen*. Version numbers like 4.0 can't be trademarked, but "USB4" can, so if a device is mislabeled they can take legal action against that.
Idea for a future tech quicky, out of order execution. Love the content keep it up!
next version will be usb 4.1 Gen 9.2 rev 7
USB 10S gen 2, 2x4 megaspeed with Greg technology, version 3 revision 17A but every spec is completely optional except for specs equivalent to usb 1.0
Once they get to a point where you can plug in portable USB graphics cards, it will open so many opportunities, especially since laptops aren't upgradable for the most part.
Riley, if you were to rename all the available USB options. What would you name them? Simplify this for everyone 🤣
USB XP
USB Vista
USB 7
USB 8
USB 8.1
USB 10
USB Phone 7 Series
@@Big_Tex Glad someone in this space has some humor 😂 it seems like the goal of every company is to see if they can cause some confusion with their naming schemes
USB
USB 2
USB 2 Episode 1
USB 2 Episode 2
USB Alyx
USB 1.0
USB 2.0
USB 360
USB One
USB One S/X
USB Series S/X
The USB
USB
Super USB
USB 64
USB W
USB U
U-Switch-B
I wonder what colour it will be? Will it be Blue like 3.0 so we can tell that we can't tell?
USB being confusing as usual, huh?
If usb name is incremental, we might have usb 11 by now since there's 8 generations of usb 3
The USB naming issues have long been obnoxious, but with USB4, they are trying to establish a better convention and this video could have educated people instead of adding to the confusion.
Version 1.0 of the USB4 specification defines a series of ports, including USB4 20Gbps and USB4 40Gbps.
Version 2.0 of the USB4 specification adds an additional port: USB4 80Gbps.
There are internal names based on how each mode works, such as USB4 2x2 and USB4 3x2 for 20 and 40 Gbps respectively. The spec also defines some other variations, as well as carries forward the definitions of the 3.2 spec.
On the whole, we would all benefit from referring to the ports by their speed going forward and should stop referring to them by the version number of the specification.
Another way to easily increase the speed is using PCIe Gen 4 (probably they are using this), the way Thunderbolt 3 and USB4 archive 40gpbs is using 4 lanes of PCIe 3.0 giving them around 32gbps (which by the way is the actual amount of speed you could use to transfer data in like an eGPU or everything that uses the PCIe lanes, In fact I'm not sure you can use those 40gbps for storage) so using 4 lanes of PCIe 4.0 would leave them almost all the way there, in fact should be easy to upgrade the USB or Thunderbolt standard to over 100gbps using PCIe 5.0 lanes, I assume they didn't go offering that due to power consumption and cost (reasons why laptops are still beginning to adopt PCIe 4.0, suspiciously good timing to implement "USB 4.1" speeds) probably also the thing that prevents them to bump up the speed following PCIe standard is getting that speed over some distance with a USB-C cable
My favorite host
It has *always* taken (many) years for a standard to finalize, and more years for it to actually come to market.
Claim your “here within an hour” ticket right here ❤️
Why are we allowing them to come up with the names though?
first
U are first but no one cares
Hello there.
idc but hi.
Correct me if I'm wrong, but those throughput speeds are per controller, not per device. That's something a lot of people forget.
1 cable for everything sounds awesome. I'm not sure why we still need Ethernet, display port, HDMI, 3.5mm, and coax? Ethernet and coax being the strangest ones.
0:25 hey bro gigabits is written Gb
Gbps
@@mikevillasenor8429 did i say gigabits or gigabits per second
This video was uploaded an hour ago. And the giveaway event from the sponsor ended 2 months ago? 3:15
This USB4v2 80Gbps speed is about the USB fabric on which different protocols (e.g. PCIe, DisplayPort, etc) can hop.
Honestly the the thing I'm looking forward to is the idea that a laptop can be your primary gaming device and when you get home have 90+ percent of the GPU, keyboard, and mouse ready to go on a single dock cable.
Going back and watching older videos and Riley’s on screen person has come a LONG way. Aging like fine wine.
Narration said gigabit per second but the graphic displayed gigabyte per second at 0:29.
OMG they can simply solve both problems by just simply having different levels meant for different things:
4.1 Minimum requirements aimed for phones
4.2 Higher requirements aimed for tablets or notebooks
4.3 Mid level requirements aimed for Laptops and consoles
4.4 High level requirements meant for powerful laptops or desktops
4.5 Max requirements aimed for top end laptops or powerful desktops
Then when someone advertises 4.4 features but only have 4.2 features, you will actually this time be able to tell the difference.
I didn't put much thought into the specifics but you get the idea
I just love the panoply of international surnames of the production team in the end credits.
1:24 Isn't it similar to qubits?
"B" is BYTE. "b" is BIT.
Finally we can charge our laptop, phones and EV with one cable :)
Big B is Bytes, little b is bits
And besides the one cable ideal, wouldn't it be nice if we could backup an internal M.2 at 10 GB/s?
This is why more than one USB 4 ports eventually are desirable as well.
3:10
"TeamGroup is currently giving away"
and the page shows that the event ended July 31st :/
Please use GB/s for GIGA BYTES per Second and Gb/s for Giga Bit / Second. Please do not use GB/s for Gigabit/s
In the near future: "USB5 Gen 4x2 v3.3+ has enough bandwidth to say its name in under a nanosecond!"
So I've been checking online for a USB C splitter that can connect a 3.5mm Jack and a USB C. And from what I've found they have limited compatibility with some phones or has some variation (18w, 30w, 60w charge, no data transfer etc) so why is this?
I thought buying a USB C at this point is a no brainer, but now you have to sift through those variations just so you can buy the right best working one.
I am thinking of a future where pcs just have like 50 type c ports on the back and you can connect any device into any port and it'll just detect what it needs to do
Except for the dedicated GPU display outputs, they could be type c but will need to plug displays into them rather than "any" of the ports
I want to buy a laptop with thunderbolt 4 for egpu gaming, should I wait to thunderbolt 5?
maybe u guys should make this kind usb video every month to keep up
Do you even realize how good of a host you are? Honestly, amazing host! Always great fun watching u!