I just found out that new motherboards are coming equipped with usb 3.1 ports. Will a mouse that is plugged into a usb 3.1 port perform better then a mouse that is plugged into a usb 2.0 or usb 3.0 port ?
orndorf77
Quote:
so there is no noticeable difference ?Originally Posted by BradleyW
In the grand scheme of things, it makes no difference.
USB 3 ports, even with USB 3 disabled in the BIOS, are much worse than USB 2 on every motherboard I've used.
In computing, there's a common terminology to define things like: multitask system, low latency system, soft real time system, hard real time system, etc.
The general rule discussing this terminology, is that there's an inverse relationship between throughput and latency. USB 2 vs USB 3 seems to follow this rule. USB 3 ports clearly have more buffering and are less responsive in all cases put out so far.
So for your question, no, there is absolutely no reason they should be the same. If anything, you might get better mouse movement on a USB 1 port.
In computing, there's a common terminology to define things like: multitask system, low latency system, soft real time system, hard real time system, etc.
The general rule discussing this terminology, is that there's an inverse relationship between throughput and latency. USB 2 vs USB 3 seems to follow this rule. USB 3 ports clearly have more buffering and are less responsive in all cases put out so far.
So for your question, no, there is absolutely no reason they should be the same. If anything, you might get better mouse movement on a USB 1 port.
Quote:
so what is the benefit of having usb 3.1 ?Originally Posted by r0ach
USB 3 ports, even with USB 3 disabled in the BIOS, are much worse than USB 2 on every motherboard I've used.
In computing, there's a common terminology to define things like: multitask system, low latency system, soft real time system, hard real time system, etc.
The general rule discussing this terminology, is that there's an inverse relationship between throughput and latency. USB 2 vs USB 3 seems to follow this rule. USB 3 ports clearly have more buffering and are less responsive in all cases put out so far.
So for your question, no, there is absolutely no reason they should be the same. If anything, you might get better mouse movement on a USB 1 port.
Quote:
Quote:
In terms of mouse movement, no you won't notice any difference unless you are a robot, or your motherboard has a terrible USB 3.0 implementation (not the case on X99).Originally Posted by orndorf77
so there is no noticeable difference ?
Quote:
More bandwidth for data transfer.Originally Posted by orndorf77
so what is the benefit of having usb 3.1 ?
Quote:
The Skylit mouse database became self aware and initiated judgement day. It then created the Skylake chipset and dropped USB 2 support so no humans will be able to properly use computer input devices and will be easier to defeat.Originally Posted by orndorf77
so what is the benefit of having usb 3.1 ?
Quote:
amd is coming out with skybridge...there is hope!Originally Posted by r0ach
The Skylit mouse database became self aware and initiated judgement day. It then created the Skylake chipset and dropped USB 2 support so no humans will be able to properly use computer input devices and will be easier to defeat.
Unless your mouse is designed for USB 3.0 or 3.1, you will NOT notice a difference over 2.0. At all.
Also, to answer your question of what the point is of having USB 3.1: USB ports are not just for mice. Think about all the different kinds of things that you can connect to a USB port, things that would benefit from a faster connection like external USB drives and whatnot.
Also, to answer your question of what the point is of having USB 3.1: USB ports are not just for mice. Think about all the different kinds of things that you can connect to a USB port, things that would benefit from a faster connection like external USB drives and whatnot.
Quote:
Can you comment on the X99 chipset?Originally Posted by r0ach
The Skylit mouse database became self aware and initiated judgement day. It then created the Skylake chipset and dropped USB 2 support so no humans will be able to properly use computer input devices and will be easier to defeat.
Quote:
Software = Drivers are more refined for high performance and stability as we expect with our 2.0 ports.
Hardware = No longer takes it's bandwidth from the PCI-E lanes, leaving PCI-E devices free from having to share resources.
I can. It has better USB 3.0 support in terms of hardware and software. I will explain in short.Originally Posted by TK421
Can you comment on the X99 chipset?
Software = Drivers are more refined for high performance and stability as we expect with our 2.0 ports.
Hardware = No longer takes it's bandwidth from the PCI-E lanes, leaving PCI-E devices free from having to share resources.
Quote:
So because X99 doesn't share resources with the gpu while using mouse, it's more responsive than Z97/87/77?Originally Posted by BradleyW
I can. It has better USB 3.0 support in terms of hardware and software. I will explain in short.
Software = Drivers are more refined for high performance and stability as we expect with our 2.0 ports.
Hardware = No longer takes it's bandwidth from the PCI-E lanes, leaving PCI-E devices free from having to share resources.
Quote:
USB 3.0 is designed to transfer data from one device to another at a faster rate.
The difference in latency is probably negligible since both platforms still require a 3rd party controller for USB 3.0.Originally Posted by TK421
So because X99 doesn't share resources with the gpu while using mouse, it's more responsive than Z97/87/77?
USB 3.0 is designed to transfer data from one device to another at a faster rate.
Quote:
Quote:
According to this article and a few others, they actually improved latency in USB 3.0:Originally Posted by r0ach
USB 3 ports, even with USB 3 disabled in the BIOS, are much worse than USB 2 on every motherboard I've used.
In computing, there's a common terminology to define things like: multitask system, low latency system, soft real time system, hard real time system, etc.
The general rule discussing this terminology, is that there's an inverse relationship between throughput and latency. USB 2 vs USB 3 seems to follow this rule. USB 3 ports clearly have more buffering and are less responsive in all cases put out so far.
So for your question, no, there is absolutely no reason they should be the same. If anything, you might get better mouse movement on a USB 1 port.
Quote:
At any rate, I would (perhaps naively) expect the changes in latency to be negligible when we're talking about mouse responsiveness on the millisecond scale, especially considering all of the other components that play a part.
Quote:
now see why nobody takes you seriously??Originally Posted by r0ach
The Skylit mouse database became self aware and initiated judgement day. It then created the Skylake chipset and dropped USB 2 support so no humans will be able to properly use computer input devices and will be easier to defeat.
If the device does not support the 3.0 protocol (no mouse to date does) it doesn't matter where you plug it in, every single HUB including 3.0 will direct the device to be hosted on a compatible controller. 1.1/2.0 devices will not be serviced with the 3.0 controller, and if they are it's because the 3.0 chip can emulate 2.0 and below behaviour. Everything will be the exact same there.
The differences that many, possibly rightly so, are skeptical about are interfacing (onboard PCI bridges vs. third-party PCI-E) and drivers/firmwares. Those may introduce some latency (data pathing, interrupt priorities, ressource sharing) or other problems with bad drivers or firmware (DPC spikes, inconsistency in regards to which host controller a device is routed to or certain protocol specifications upon booting or plugging).
3.0 mainly was supposed to increase transmission bandwidth and power supply with things like tablets and smartphones in mind, I'm not sure polled external buffers like with mice can be accessed more rapidly with 3.0, but if it is possible it comes down to manufacturers specifically modifying mice to support the 3.0 communication and they won't do that until 3.0 is the standard. Going 3.0 now denies your product a big chunk of market attraction.
Until there's true 3.0+ support with mice, there's no reason to plug them into anything but <2.0 ports. Best case scenario everything is the exact same, worse case there's latency or inconsistency, worst case you blow up your mouse (I once fried an USB sound card with 3.0 becuse of compatibility/driver issues resulting in overvolting).
The differences that many, possibly rightly so, are skeptical about are interfacing (onboard PCI bridges vs. third-party PCI-E) and drivers/firmwares. Those may introduce some latency (data pathing, interrupt priorities, ressource sharing) or other problems with bad drivers or firmware (DPC spikes, inconsistency in regards to which host controller a device is routed to or certain protocol specifications upon booting or plugging).
3.0 mainly was supposed to increase transmission bandwidth and power supply with things like tablets and smartphones in mind, I'm not sure polled external buffers like with mice can be accessed more rapidly with 3.0, but if it is possible it comes down to manufacturers specifically modifying mice to support the 3.0 communication and they won't do that until 3.0 is the standard. Going 3.0 now denies your product a big chunk of market attraction.
Until there's true 3.0+ support with mice, there's no reason to plug them into anything but <2.0 ports. Best case scenario everything is the exact same, worse case there's latency or inconsistency, worst case you blow up your mouse (I once fried an USB sound card with 3.0 becuse of compatibility/driver issues resulting in overvolting).
Quote:
If I have USB 3 disabled in the BIOS and plug a mouse into the USB 3 port, best case scenario, I'm getting a pass through mode to the USB 2 controller. Most devices operating in pass through mode have latency problems. USB controllers are mega intensive IO devices, so there's no reason you would want two enabled at the same time. Since there's no USB 3 mice, disabling USB 3 and plugging things only into the USB 2 port is the most logical option.
If Skylake removes USB 2 and I plug in a USB 2 mouse, instead of operating in pass through mode, I'm going to be operating in legacy mode on the actual USB 3 controller. In this scenario, it's possible it could feel the same as USB 2, and it's possible it would inherit more buffering and such from the higher throughput of USB3 devices. It could go either way and vary by each Intel chipset release.
It's not hard to see why USB 3 would be laggier than USB 2 right now:Originally Posted by HAGGARD
If the device does not support the 3.0 protocol (no mouse to date does) it doesn't matter where you plug it in, every single HUB including 3.0 will direct the device to be hosted on a compatible controller. 1.1/2.0 devices will not be serviced with the 3.0 controller, and if they are it's because the 3.0 chip can emulate 2.0 and below behaviour.
If I have USB 3 disabled in the BIOS and plug a mouse into the USB 3 port, best case scenario, I'm getting a pass through mode to the USB 2 controller. Most devices operating in pass through mode have latency problems. USB controllers are mega intensive IO devices, so there's no reason you would want two enabled at the same time. Since there's no USB 3 mice, disabling USB 3 and plugging things only into the USB 2 port is the most logical option.
If Skylake removes USB 2 and I plug in a USB 2 mouse, instead of operating in pass through mode, I'm going to be operating in legacy mode on the actual USB 3 controller. In this scenario, it's possible it could feel the same as USB 2, and it's possible it would inherit more buffering and such from the higher throughput of USB3 devices. It could go either way and vary by each Intel chipset release.
Disabling the USB 3.0 ports makes no difference to how USB 2.0 behaves on a noticeable level.
Well, as soon as the respective chip has control over the device "passthrough" is really just electrical signals taking a slightly different route on the motherboard. But as I mentioned there, they may have to share that route with other devices more so than they would on a properly implemented onboard hub.
Assuming 3.0 is equally properly implemented in the mobo as standard "then-legacy" ports and doesn't come with bad drivers/firmwares, I really can't see how the behaviour wouldn't be the exact same. But you kind of said that as well.
I agree at least that disabling third-party 3.X controllers can have positive effects. Or as many USB controllers as are not currently occupied for that matter. As a general mentality I do support "disbling stuff you don't use". I personally don't use 3.0 on my desktop, in general I just delegate most of my multimedia needs to my notebook (including printing, swap media, smartphone support, camera, optical drive, etc.) allowing me to disable/disconnect all that jazz on the PC I'm gaming on for good measure. That doesn't mean I think there's a direct difference to mouse behaviour with those things, but the general system (DPC, polling precision, power consumption) might get a tad bit snappier. I have also tested just about every HPET combination and while I know it's somewhat dependent on hardware, I haven't noticed a difference in mouse feel with either, effects on DPC and polling are undoubtedly there though.
Assuming 3.0 is equally properly implemented in the mobo as standard "then-legacy" ports and doesn't come with bad drivers/firmwares, I really can't see how the behaviour wouldn't be the exact same. But you kind of said that as well.
I agree at least that disabling third-party 3.X controllers can have positive effects. Or as many USB controllers as are not currently occupied for that matter. As a general mentality I do support "disbling stuff you don't use". I personally don't use 3.0 on my desktop, in general I just delegate most of my multimedia needs to my notebook (including printing, swap media, smartphone support, camera, optical drive, etc.) allowing me to disable/disconnect all that jazz on the PC I'm gaming on for good measure. That doesn't mean I think there's a direct difference to mouse behaviour with those things, but the general system (DPC, polling precision, power consumption) might get a tad bit snappier. I have also tested just about every HPET combination and while I know it's somewhat dependent on hardware, I haven't noticed a difference in mouse feel with either, effects on DPC and polling are undoubtedly there though.
Quote:
It's pretty easy to notice the difference between having 2 USB controllers running and only 1 running. The USB controller is like the most intensive IO device on the entire system. If settings like HPET are acknowledged as changing mouse movement, then adding a second USB controller is obviously going to do something. It's like plugging in a second sound card, it's not some minor thing.Originally Posted by BradleyW
Disabling the USB 3.0 ports makes no difference to how USB 2.0 behaves on a noticeable level.
Quote:
If nothing is plugged into the USB 3.0 ports, it won't change how USB 2.0 behaves. If the application you are running takes advantage of HPET, only then will it have some kind of effect on the system. Most applications will use the OS default timer which is usually the LAPTIC or TSC depending on your primary platform.Originally Posted by r0ach
It's pretty easy to notice the difference between having 2 USB controllers running and only 1 running. The USB controller is like the most intensive IO device on the entire system. If settings like HPET are acknowledged as changing mouse movement, then adding a second USB controller is obviously going to do something. It's like plugging in a second sound card, it's not some minor thing.
Quote:
We already know both of those statements are false. At the very least, the USB controller is going to take an IRQ and other system resources even if nothing is plugged into it. Even though people pretend that IRQ allocatoin doesn't matter anymore, people are posting things on this forum all the time that show otherwise. That's just a side issue though because the USB controller does more than just occupy IRQs.
The overwhelming majority of this forum also notices a difference in HPET on and off in the BIOS regardless of what the OS is doing with it. Trying to claim HPET does nothing is just absurd. Even the people on Guru3d for god's sake that use 4 video cards at the same time and don't optimize their system for latency at all notice that HPET on/off makes a big difference in mouse movement.
......Originally Posted by BradleyW
If nothing is plugged into the USB 3.0 ports, it won't change how USB 2.0 behaves. If the application you are running takes advantage of HPET, only then will it have some kind of effect on the system. Most applications will use the OS default timer which is usually the LAPTIC or TSC depending on your primary platform.
We already know both of those statements are false. At the very least, the USB controller is going to take an IRQ and other system resources even if nothing is plugged into it. Even though people pretend that IRQ allocatoin doesn't matter anymore, people are posting things on this forum all the time that show otherwise. That's just a side issue though because the USB controller does more than just occupy IRQs.
The overwhelming majority of this forum also notices a difference in HPET on and off in the BIOS regardless of what the OS is doing with it. Trying to claim HPET does nothing is just absurd. Even the people on Guru3d for god's sake that use 4 video cards at the same time and don't optimize their system for latency at all notice that HPET on/off makes a big difference in mouse movement.
Quote:
As for HPET, its about equal on the "for" and "against". Either it's placebo or it's platform specific. Makes no difference on my X79 platform. HPET option is also forced ON in X99.
If you tell someone that disabling a feature will make things better, there is a good chance they will perceive a difference regardless.
With the early implementation of USB 3.0, yes it did take up some system resources and added slight latency, but nothing on a perceivable level. On the X99 platform, USB 3.0 is implemented properly, so both latency and resource usage have been improved to the point that its not even worth mentioning.Originally Posted by r0ach
......
We already know both of those statements are false. At the very least, the USB controller is going to take an IRQ and other system resources even if nothing is plugged into it. Even though people pretend that IRQ allocatoin doesn't matter anymore, people are posting things on this forum all the time that show otherwise. That's just a side issue though because the USB controller does more than just occupy IRQs.
The overwhelming majority of this forum also notices a difference in HPET on and off in the BIOS regardless of what the OS is doing with it. Trying to claim HPET does nothing is just absurd. Even the people on Guru3d for god's sake that use 4 video cards at the same time and don't optimize their system for latency at all notice that HPET on/off makes a big difference in mouse movement.
As for HPET, its about equal on the "for" and "against". Either it's placebo or it's platform specific. Makes no difference on my X79 platform. HPET option is also forced ON in X99.
If you tell someone that disabling a feature will make things better, there is a good chance they will perceive a difference regardless.
Quote:
And sorry to burst your bubble but we're not having mass hysteria here, it does change how the system feels in a perceivable manner on a lot of the platforms. The reason you're not perceiving difference is probably because your system is already bottlenecked/too bloated with different USB devices, IRQ sharing and other BIOS settings and software.
I haven't tested X79 or X99 but I'd like comments from people who can feel the difference.
First of, you're such a hypocrite with this statement.Originally Posted by BradleyW
With the early implementation of USB 3.0, yes it did take up some system resources and added slight latency, but nothing on a perceivable level. On the X99 platform, USB 3.0 is implemented properly, so both latency and resource usage have been improved to the point that its not even worth mentioning.
As for HPET, its about equal on the "for" and "against". Either it's placebo or it's platform specific. Makes no difference on my X79 platform. HPET option is also forced ON in X99.
If you tell someone that disabling a feature will make things better, there is a good chance they will perceive a difference regardless.
And sorry to burst your bubble but we're not having mass hysteria here, it does change how the system feels in a perceivable manner on a lot of the platforms. The reason you're not perceiving difference is probably because your system is already bottlenecked/too bloated with different USB devices, IRQ sharing and other BIOS settings and software.
I haven't tested X79 or X99 but I'd like comments from people who can feel the difference.
-
?
-
?
-
?
-
?
-
?
-
?
-
?
-
?
-
?
-
?
-
?
-
?
-
?
-
?
-
?
-
?
-
?
-
?
-
?
-
?
