Does the USB port matter for Keyboard/Mouse?

[fec32a4de]

Hello

Got an interesting, perhaps stupid question for yall....

Does it matter which USB port you use for a keyboard or a mouse?

I got an X570 mobo with 4 USB2.0 ports, a few 3.0 (5Gbps ports) and 3.1 Gen 1 Service Pack 3 Professional lite edition whateveritscallednow (10Gbps ports)

The connectors on every any mouse and keyboard I have in the house has 4 pins for USB2.0, cause they don't need more bandwidth but whatever.

So: does it actually matter if I use 480Mbps ports, 5Gbps ports or 10Gbps ports for keyboard and mouse? I don't have a lot of USB devices so I got free ports to go around.

Maybe some are CPU attached vs PCH/Chipset attached? Block diagram isn't available, but does it really make a difference?

Microseconds, few milliseconds?

Thanks

 

[P4-630]

which USB port you use for a keyboard or a mouse?

Just use USB 2.0 ports.

 

[Bill_Bright]

Got an interesting, perhaps stupid question for yall....

Does it matter which USB port you use for a keyboard or a mouse?

No.

And FTR, you can use a USB 3.x port. It won't hurt anything and it will work just fine. However, it is common for most systems to have both USB 2.x and USB 3.x ports. Since keyboards and mice do not use or require the extra bandwidth/speeds USB 3.x support, many folks recommend sticking with the USB 2.x ports for keyboards and mice to ensure there are enough USB 3.x ports available for the devices that do take advantage of the higher speeds.

Another factor to consider is power through the ports. Many newer motherboards allow wireless devices to be charged when connected to designated USB ports, even when in standby/sleep modes. This is a common feature for many laptops. That is, when the laptop is in sleep mode but connected to the charger, you can charge your cell phone, for example, through one of those USB power ports. Again, check your manual. You typically want to leave those power ports open so you can charge other devices, and connect the keyboard and mouse to different ports.

 

[ThrashZone]

Hi,
Use the higher speed usb ports for high speed devices like flash drives/...
Mouse and keyboard just look at the tips on them they are likely black meaning usb 2.0 so use that port on the board and sometime usb devices react differently on faster usb ports and not in a good way hehe

 

[mojeziomkitopieniazki]

It matters. Basically, like the guy above said: just look at the color of the USB port and go with that. But from my experience, mice and keyboards don’t work well with newer USB ports.

Just like that:
2.0 - mouse and keyboard, best stability and compatibility
3.x - fastest, meant for stuff like hard drives and flash drives

 

[eidairaman1]

Doesnt matter, just plug and go

 

[DaemonForce]

It kind of does on some X570 boards. I've noticed that the top USB slots work best for my sensors and the middle USB is finnicky.
Bottom of the I/O plate is super stable but has issues with some VR sensors at times.
I picked up a cheap USB header bracket and use that to plug in my mice and keyboards just to save USB ports.
I also have splitters but the chips associated with each USB definitely seem to make a difference.

 

[Ruru]

Just use USB 2.0 ports.

Agree. I've always used the upmost USB2.0 ports for KB and mouse.

edit: Doesn't keyboards and mice itself use USB1.1 or even 1.0 IIRC?

 

[lexluthermiester]

Hello

Got an interesting, perhaps stupid question for yall....

Does it matter which USB port you use for a keyboard or a mouse?

I got an X570 mobo with 4 USB2.0 ports, a few 3.0 (5Gbps ports) and 3.1 Gen 1 Service Pack 3 Professional lite edition whateveritscallednow (10Gbps ports)

The connectors on every any mouse and keyboard I have in the house has 4 pins for USB2.0, cause they don't need more bandwidth but whatever.

So: does it actually matter if I use 480Mbps ports, 5Gbps ports or 10Gbps ports for keyboard and mouse? I don't have a lot of USB devices so I got free ports to go around.

Maybe some are CPU attached vs PCH/Chipset attached? Block diagram isn't available, but does it really make a difference?

Microseconds, few milliseconds?

Thanks

Yes and no. Ideally, if you have USB2.0 ports, use them for your KB&M & gamepads. USB2.0 is way more than enough for USB human input devices.

edit: Doesn't keyboards and mice itself use USB1.1 or even 1.0 IIRC?

Yes, they did and they were fine.

 

[Onasi]

The only exceptions to the rule of “just use USB 2.0” are the high-polling mice and keyboards. By that I mean those above 1000 Hz. This is due to faster interrupts that XHCI can provide and using that by default necessitates USB 3.x. It’s a very niche use-case, granted. And no, regular 1000 Hz devices will not benefit in any way.

 

[Dr. Dro]

You can plug them anywhere, although it's preferable for you to use a low-bandwidth port whenever possible, as to avoid wasting system resources. USB 1.1 ports are enough, and some systems (late 2000s to mid 2010s Mac computers in particular) actually tend to prefer and will establish an USB 1.1 link when the detected device is a keyboard or mouse, allegedly to save power.

Some newer computers no longer include USB 1.1 or 2.0 controller hubs. In these cases, always choose the slowest available ports for these peripherals. PS/2 ports are still fine, although it's harder to find devices compatible with them - and most motherboards no longer come with a PS/2 port at all. Usually it's a hybrid keyboard/mouse input port behind an embedded converter, meaning you can only use one or the other and only on modern OSes.

Motherboards intended for extreme overclocking are amongst the few to still include native PS/2 ports, presumably because some benchmarks run better under Windows XP, making them necessary.

It kind of does on some X570 boards. I've noticed that the top USB slots work best for my sensors and the middle USB is finnicky.
Bottom of the I/O plate is super stable but has issues with some VR sensors at times.
I picked up a cheap USB header bracket and use that to plug in my mice and keyboards just to save USB ports.
I also have splitters but the chips associated with each USB definitely seem to make a difference.

USB on Zen 3-based systems is a mess. On client computers, AGESA in the 1.2.0.0 to 1.2.0.5 range is susceptible to sudden crashes and stack resets - and similar to an errata that AMD has refused to fix for EPYC Milan processors that states that the USB host controller may crash and reset at any time. Supposedly operating the bus in a certain pattern works around the problem and AMD's basically told operating system vendors to deal with it as they see fit.

 

[Ruru]

USB on Zen 3-based systems is a mess. On client computers, AGESA in the 1.2.0.0 to 1.2.0.5 range is susceptible to sudden crashes and stack resets - and similar to an errata that AMD has refused to fix for EPYC Milan processors that states that the USB host controller may crash and reset at any time. Supposedly operating the bus in a certain pattern works around the problem and AMD's basically told operating system vendors to deal with it as they see fit.

I've only had minor problems with my old Oculus Rift CV1 with my B550 board, and they were with R5 3600. Actually upgrading to Zen3 squashed those USB bugs, at least with my PC.

Updated my BIOS also yesterday and on changelog Gigabyte says "Update AMD AGESA V2 1.2.0.C", if that means anything. I'm not that familiar with AGESA and what it even actually does.

 

[Vayra86]

It matters because you will never need a chipset driver installed for your USB 2.0 ports on the rear I/O plate. They work out of the box and neither do you need to hook up any internal 3.0 plug to the case. That plug can fail as well. Beyond that, its just about efficient use of your ports for other stuff as mentioned.

 

[Bill_Bright]

It depends.

Technically the mouse and keyboard "should" work via any USB port. However, some motherboards have designated ports for the mouse and keyboard. As Vayra86 noted above, they are located in the rear panel I/O area. They are "special" because they (at least the keyboard port) work at the basic hardware level - meaning you can boot into the BIOS Setup Menu (or into Safe Mode) and be certain the keyboard (and probably mouse too) will be properly recognized so you can navigate around the menu, or Windows in Safe Mode.

Remember, when booting into the BIOS Setup Menu, this happens BEFORE the boot drive is touched and therefore BEFORE any special drivers (including USB hub drivers) are loaded. And in Safe Mode, only basic drivers are loaded. All keyboards and mice, as well as all motherboards (and operating systems, graphics solutions and monitors too) are designed to work at "basic" levels without any additional drivers. If not, how could we "see" and complete OS installations without keyboard function or a display on our monitors before installing drivers?

So again, it depends - on the motherboard but also note this information is typically clearly printed in the motherboard's user manual.

I will say I always use the designated keyboard and mouse ports simply because it makes it easier to "intuitively" know which cable goes where when staring at a dusty rat's nest of cables.

 

[dgianstefani]

I use 8000 Hz hyperpolling, so I ensure to connect the transceiver to a USB 3.x port directly routed to the CPU, not through a chipset first.

For 1000 Hz and below, it does not matter, unless the peripheral requires more wattage than the USB 2.0 port can provide, which is unlikely.

Check the relevant section in any high polling mouse review.

MCHOSE AX5 Pro Max Review

Weighing 55 g, the ambidextrous AX5 Pro Max comes with a magnesium alloy shell. Equipped with PixArt's PAW3395 sensor, the AX5 Pro Max is capable of both Bluetooth and 2.4 GHz, achieving up to 130 hours of battery life, along with 8000 Hz wired and wireless polling.

www.techpowerup.com

In order to get the full benefit out of 8000 Hz polling, certain conditions need to be met. First, it is recommended to have a sufficiently powerful CPU; i.e., one with six physical cores and appropriately high IPC at the least. Second, the OS has to be capable of interrupt moderation of 125 μs or lower. This is true of Windows 8 or higher, where interrupt moderation on XHCI will typically be 50 μs, but not of Windows 7 and lower, where interrupt moderation is never below 1 ms unless changed manually, which isn't easily done. On EHCI, interrupt moderation can be expected to be 125 μs on Windows 8 or higher, which is sufficient but not optimal. Third, it is therefore recommended to plug the AX5 Pro Max into a USB 3.x port in XHCI mode. Any USB 3.x ports forced into EHCI will behave similarly to a native USB 2.0 port. As a general rule of thumb, one should be using a USB port native to the CPU and not connect any other high-polling devices to a port of the same hub. Even if all of these conditions are met, actual polling stability during higher workloads will further depend on general system and OS health. As such, it is recommended to use a reasonably optimized OS installation without bloat in conjunction with the AX5 Pro Max.

If you want sub 1 ms click latency along with the other benefits, such as better tracking etc, then you're unlikely to get that from a 1000 Hz or lower mouse.

Therefore, use a native USB 3.0 port if your device is fast enough to use it.

 

[Onasi]

If you want sub 1 ms click latency along with the other benefits, such as better tracking etc, then you're unlikely to get that from a 1000 Hz or lower mouse.

Nah, the click latency is in most cases switch implementation and firmware related. The current champion is the OP1 8K:

And, as pzogel notes himself in the review:
“Click latency has been measured to be 0.1 (0.12) ms, with standard deviation being 0.00 ms. Polling rate has no effect on click latency.”
And Speed Mode is solely physical in its nature, so again, polling won’t matter there.

Tracking improvements ARE there up to, debatably, either 2000 or 4000Hz, provided one has a very high refresh screen. The diminishing returns are insane in those, however. Even then, a lot of people in the aiming community consider said benefits to be marginal at the very best, so YMMV.

I guess what I am saying here is that while it’s true that there are theoretical and measurable benefits to high-polling input devices and running them off the USB 3.x port is good practice, how really practically relevant those are compared to a good 1000Hz device plugged into a USB 2.0 port is… rather questionable. I, personally, maintain that the current high-polling craze is the “DPI wars” of yesteryear for mouse manufacturers, especially considering that most professional players haven’t switched and seem to be doing just fine.

 

[dgianstefani]

Nah, the click latency is in most cases switch implementation and firmware related. The current champion is the OP1 8K:
And, as pzogel notes himself in the review:
“Click latency has been measured to be 0.1 (0.12) ms, with standard deviation being 0.00 ms. Polling rate has no effect on click latency.”
And Speed Mode is solely physical in its nature, so again, polling won’t matter there.

On that specific wired mouse polling rate has no effect on click latency.

Tracking improvements ARE there up to, debatably, either 2000 or 4000Hz, provided one has a very high refresh screen. The diminishing returns are insane in those, however. Even then, a lot of people in the aiming community consider said benefits to be marginal at the very best, so YMMV.

I guess what I am saying here is that while it’s true that there are theoretical and measurable benefits to high-polling input devices and running them off the USB 3.x port is good practice, how really practically relevant those are compared to a good 1000Hz device plugged into a USB 2.0 port is… rather questionable. I, personally, maintain that the current high-polling craze is the “DPI wars” of yesteryear for mouse manufacturers, especially considering that most professional players haven’t switched and seem to be doing just fine.

Would you be able to explain to me how 0.1 ms click latency could be achieved with 1000 Hz, which refreshes every 1 ms?

Many mice have improved click latency as polling rate increases. Some do not, but I suspect this is because clicks are transmitted at a rate higher than 1000 Hz, i.e. whenever they are made, but tracking is not.

For example the Viper V3 Hyperspeed, which goes from 0.8, 0.6, 0.5, 0.45 ms as polling rate increases.

This kind of "flexible" polling rate is also seen in other respects, where the mouse will only actually poll at those higher rates if the CPI is set sufficiently high, otherwise it will poll as high as the CPI allows for, to save battery.

It also seems pretty clear to me that the click latency charts are dominated by high polling rate mice. Whether that is unrelated or related is open to interpretation, but I can certainly see the advantages of sub ms polling if you are looking to minimize latency.

Set to 8000 Hz, motion delay is lowered by roughly 1 ms, bringing the Viper V3 Hyperspeed 0.5 ms ahead of the G403 instead of behind it, and click latency is lowered to just 0.45 ms

However, it is important to note that those values merely denote the maximum applicable polling rate. If the mouse isn't physically moved enough to generate a sufficient number of motion events (for 8000 Hz at least 8000 pixels worth of motion per second), less updates will be transmitted, resulting in a lower effective polling rate. Accordingly, it is strongly recommended to use a sufficiently high CPI step on the Viper 8K.

 

[Dr. Dro]

I've only had minor problems with my old Oculus Rift CV1 with my B550 board, and they were with R5 3600. Actually upgrading to Zen3 squashed those USB bugs, at least with my PC.

Updated my BIOS also yesterday and on changelog Gigabyte says "Update AMD AGESA V2 1.2.0.C", if that means anything. I'm not that familiar with AGESA and what it even actually does.

Yeah on AM4 those problems hit Zen 2 too. I had a ROG Strix B550-E and had constant reset issues with all 3 Ryzen CPUs I owned at the time, 3500X, 3900XT and 5950X. I was one of the people that AMD contacted on Reddit at the time for diagnostic data, which I promptly provided. Those months really sucked

 

[Ruru]

Yeah on AM4 those problems hit Zen 2 too. I had a ROG Strix B550-E and had constant reset issues with all 3 Ryzen CPUs I owned at the time, 3500X, 3900XT and 5950X. I was one of the people that AMD contacted on Reddit at the time for diagnostic data, which I promptly provided. Those months really sucked

I remember those times, luckily I had problems only with the VR headset and I use that hella rare anyway so personally not a game breaker.

Though now I don't have even those few USB problems anymore.

 

[Dr. Dro]

It depends.

Technically the mouse and keyboard "should" work via any USB port. However, some motherboards have designated ports for the mouse and keyboard. As Vayra86 noted above, they are located in the rear panel I/O area. They are "special" because they (at least the keyboard port) work at the basic hardware level - meaning you can boot into the BIOS Setup Menu (or into Safe Mode) and be certain the keyboard (and probably mouse too) will be properly recognized so you can navigate around the menu, or Windows in Safe Mode.

Remember, when booting into the BIOS Setup Menu, this happens BEFORE the boot drive is touched and therefore BEFORE any special drivers (including USB hub drivers) are loaded. And in Safe Mode, only basic drivers are loaded. All keyboards and mice, as well as all motherboards (and operating systems, graphics solutions and monitors too) are designed to work at "basic" levels without any additional drivers. If not, how could we "see" and complete OS installations without keyboard function or a display on our monitors before installing drivers?

So again, it depends - on the motherboard but also note this information is typically clearly printed in the motherboard's user manual.

I will say I always use the designated keyboard and mouse ports simply because it makes it easier to "intuitively" know which cable goes where when staring at a dusty rat's nest of cables.

Just to add a little to this, regarding the ports it doesn't matter anymore. There is no need to use any designated port and in fact most motherboards no longer have one. To navigate in preinstallation environments, DOS and other sorts of legacy operating systems, there is a "Legacy USB mode" in most BIOS settings. Simply having this on will have the BIOS pass keyboard commands to said applications.

USB 2.0 ports on the back panel I/O were removed in newer motherboards, for example, my Z790 Apex Encore has 4 5 Gbps ports, 5 10 Gbps ports and a 20 Gbps Type-C port on the back. None of the old 2.0 connectors, just kept as an option for front I/O on older case models.

I am sooo glad to see that this wretched format heading to the fiery pits of hell from whence it came 20+ years ago. SATA and the mp3 codec should follow.

 

[Ruru]

Z790 Apex Encore

I almost spat my coffee when I saw the price of that board. I simply can't understand how it's even possible to make a board what costs 800EUR and it's not even a limited edition (like ARES, MARS etc back in the day)

Personally I'm glad that PS/2 ports have disappeared since the compatibility of USB has been good for years.

 

[Dr. Dro]

I almost spat my coffee when I saw the price of that board. I simply can't understand how it's even possible to make a board what costs 800EUR and it's not even a limited edition (like ARES, MARS etc back in the day)

Personally I'm glad that PS/2 ports have disappeared since the compatibility of USB has been good for years.

Yeah PS/2 ports have only been kept on these extreme boards for benchmarking purposes since USB support is flaky and drivers haven't been and won't be written for older OSes like XP.

It's expensive, but it's super good. I think having a solid motherboard foundation is important, so it's the one component I don't hesitate to splurge. Still, I agree it's not for everyone and I wouldn't even dare recommending it for the average user. Most people (and arguably myself) would likely be better served by a more down to earth choice at around half the price, or if still going with the blank check option, the Dark Hero or Formula variants which should be better for gaming PCs - this one is just focused at pure performance and has certain tradeoffs to match (audio is less powerful, only 2 memory slots, etc).

 

[Onasi]

Would you be able to explain to me how 0.1 ms click latency could be achieved with 1000 Hz, which refreshes every 1 ms?

Sure. 1000Hz can add up to 1ms. At maximum, not every time. Depends on when the click event happens relative to the poll. And from my understanding, the TPU test accounts for it by taking the lowest LDAT measurement. Paging @pzogel for clarification.

Many mice have improved click latency as polling rate increases. Some do not, but I suspect this is because clicks are transmitted at a rate higher than 1000 Hz, i.e. whenever they are made, but tracking is not.

For example the Viper V3 Hyperspeed, which goes from 0.8, 0.6, 0.5, 0.45 ms as polling rate increases.

This kind of "flexible" polling rate is also seen in other respects, where the mouse will only actually poll at those higher rates if the CPI is set sufficiently high, otherwise it will poll as high as the CPI allows for, to save battery.

It also seems pretty clear to me that the click latency charts are dominated by high polling rate mice. Whether that is unrelated or related is open to interpretation, but I can certainly see the advantages of sub ms polling if you are looking to minimize latency.

We’ve seen sub-1ms latency from mice when optical switches had came out. And the OG Viper and Viper Mini were 1000Hz mice. Same as the M4 Air that held a near top spot for years with conventional mechanical switches and a 1k polling rate. The latest MCHOSE mouse also shown no difference in click latency between polling rates. I genuinely think this is fully a function of the firmware and MCU implementation. After all, we also get things like these that don’t make sense… until you realize that the 8K dongle uses a different MCU.

 

[Bill_Bright]

Just to add a little to this, regarding the ports it doesn't matter anymore. There is no need to use any designated port and in fact most motherboards no longer have one.

I don't see the point of this, except to be contrary.

As I noted in the text you quoted, these devices should work in any USB port. I also noted with some motherboards, they designate one or two for those devices to ensure they do work in the BIOS Setup Menu (so no need to worry about how one would set the legacy setting).

Contrary to your implied assumption, not all motherboards are newer generation boards. As you noted, "most" no longer designate ports, implying some still do - as noted. And even some still support PS/2 - though typically today, there is just one PS/2 port that is dual function (as in mouse "or" keyboard).

 

[dgianstefani]

Sure. 1000Hz can add up to 1ms. At maximum, not every time. Depends on when the click event happens relative to the poll. And from my understanding, the TPU test accounts for it by taking the lowest LDAT measurement. Paging @pzogel for clarification.

We’ve seen sub-1ms latency from mice when optical switches had came out. And the OG Viper and Viper Mini were 1000Hz mice. Same as the M4 Air that held a near top spot for years with conventional mechanical switches and a 1k polling rate. The latest MCHOSE mouse also shown no difference in click latency between polling rates. I genuinely think this is fully a function of the firmware and MCU implementation. After all, we also get things like these that don’t make sense… until you realize that the 8K dongle uses a different MCU.

So why does the same mouse set to a higher polling rate provide better click latency?

For obvious reasons having a higher polling rate will improve consistency of low click latency, rather than simply hoping the click registers at a period of time close to when the mouse reports.