I haven't logged on here in like a year+ and came on and saw a bunch of mail saying all this new hardware is giving people terrible mouse movement asking what motherboard exists now that doesn't suck. Since at one point in time I had an entire room full of computer crap mining digital $hitcoins and just purchased one of everything because why not, I have personal experience with most brands of motherboards and GPUs in the 3570k-4570k era, and also 2500k era too. Here's what I noticed during that period.
: They were all garbage. Mouse movement was crap on every single one I've tried. One of the boards even caught on fire while I ran Intel Burn Test and sent this chemical, toxic smoke cloud through the room that made me feel sick for a while after inhaling it. Probably gave me brain damage as well as bad mouse movement. I think one even allowed you to turn off HPET in the BIOS but when I checked, it turns out the flag did absolutely nothing and it was still on. HAHA SCREW YOU.
: This is a complex issue. Back in the 2500k era, you had flagship Gigabyte motherboards like the UD5H that worked pretty well mouse-wise, but suddenly they released a new BIOS version midway through the board's life that REALLY REALLY dulled down mouse response. This was during an era of motherboard manufacturers not even having adopted UEFI yet (I use legacy BIOS exclusively), and I think Gigabyte was in the process of somewhat half ass supporting it.
From memory, I seem to recall the newer BIOS update changed the BIOS graphical interface and tossed in some more UEFI compatibility, and that's when the mouse response cratered. I haven't purchased a Gigabyte board since, so I don't know if Gigabyte boards went to crap from that point on due to some bad firmware implentation or if they've fixed things. It's been a long time so anything could have happened. Gigabyte boards BEFORE that point were some of the best DPC latency-wise, which is why they were purchased by people like audio engineers a lot.
: These have typically always been the best for me. The only problem is that they have no option to turn HPET off on many boards. They also install a bunch of proprietary power control systems and all kinds of weird crap that might have a negative effect on mouse use if they don't give you the right BIOS options to control it or turn them off. You might be better off just buying the cheapest ASUS board possible to avoid potential issues with any of their proprietary stuff that's usually not on cheaper boards to have a more reference-style system.
: While not as bad as Asrock, these have never been good for me and always inferior mouse-wise compared to Asus boards. I can't really compare them to Gigabyte boards since I don't have experience with new ones.
Why is there such a discrepency in all these boards? I think a lot of it is solely due to their BIOS implementations and it's probably not even the hardware itself. Japan has more assembly and other esoteric coders than anywhere I think, and Asus is designed in Japan so might have more competent people? Who knows, they might all be made in China now, haven't kept up. I haven't used a new AMD board, so no idea if they're terrible or not. I would not use one unless you can make it run with the default Microsoft Windows USB 2.0 controller driver instead of some 3rd party one.
Another important thing to note is that many of these motherboards will have like 10+ versions during their lifecycle and usually one version - or a few in a row - will be vastly superior to the others in terms of mouse response. They're constantly doing experimental changes and some of them WILL extremely negatively affect your mouse response like the Gigabyte example above. Case in point, the Alienware alpha box. There's eight BIOS versions and A05 is the best one, while the three after that introduce the infamous Intel CPU exploit fix that also happens to turn mouse response into a slug.
I wouldn't be surprised if EVERY Intel exploit BIOS fix released in that timeframe has a similar effect on all these manufacturers. You might even need a Skylake or earlier CPU because Intel attempting a hardware fix without creating a new architecture might cause the same sluggish response issues as the software fix. So if you really wanted to be safe, you would want to run a burner gaming box with Windows 8.1 with no updates (unless you manually want to test each and every one since they started backporting Win 10 garbage), and probably a Skylake or earlier CPU without the meltdown and spectre OS or BIOS patches (because I personally witnessed the BIOS patches cratering mouse response on the Alienware Alpha for instance).
Also, you want the earlier Win 8.1 x64 Pro build before they attempted to force you to create a Microsoft account to install it: SHA1: 9D064479070A364FDC034C7B2DD67AAA8A4AA147
And here's the elephant in the room most people don't know about but probably wish you did. Reference video cards vs 3rd party. I've had like 4 or 5 different brands of AMD R9 290 all at the same time and ALL the 3rd party, non-reference cards had inferior, wonky mouse movement compared to the reference one. I haven't had literally every single brand of Nvidia card all at the same time from a specific generation, only AMD, so I haven't tested this thorougly for Nvidia, but I think reference Nvidia are better as well. If you're a pro or just annoyed by bad mouse movement, buy only reference cards.
On the subject of video card drivers
On the subject of keyboards
Originally Posted by r0ach
Nvidia driver 441.66 seems to be the best for mouse response and cursor movement in a long time. I've tested most of them since the "ultra low latency" Nvidia setting came out, and all of them were worse than 441.66. If you're sitting on the fence waiting for a 7nm 2020 Nvidia card, I would not wait because you could end up with a long period of having no drivers like 441.66 for it. The last Nvidia driver I actually liked for mouse response before this was 347.09 beta (not the WHQL which was worse) for reference. Do you want to wait years for a new decent driver?
I'd be surprised if the 5700XT and AMD's historically buggier and less optimized drivers are better right now, but I haven't tried one. Almost all of Nvidia's default settings need to be changed. Settings like Anisotropic filtering for instance. From memory, I don't believe leaving this on default/application controlled was that detrimental in the older days. But ANYONE can easily notice how big of an issue this setting is now by loading up an Unreal Engine 4 game like Borderlands 3 that allows you to toggle between trilinear filtering and anistropic filtering - not many games nowadays allow you to use trilinear at all for some reason. Trilinear filtering is WAY less laggy on these new age Nvidia cards and drivers. Nvidia seems to allocate some type of enormous buffering to anistropic filtering, probably as a ghetto performance hack, so by setting it to off in Nvidia control panel, it seems to get rid of it.
Here's all the current Nvidia settings I use. I haven't put in a ton of time tinkering with newer settings like "OpenGL Rendering GPU", but it seemed like manually selecting the GPU instead of leaving that one on auto gave weird floaty cursor feeling. It's probably configuring something stupid with SLI in the background if you force it, making auto the 'non-SLI' choice, even though the driver doesn't tell you that. Having "Texture Filtering Quality" on the default of "Quality" even seems a bit more responsive now than manually changing it to "High Quality". So you have to be careful if you're going for optimal mouse cursor movement, and things change with various driver releases.
Important: You also need to change this setting under "Desktop and Color Settings". I don't think the setting comes up if you connect by DVI, but does for HDMI and possibly others:
Haven't logged on in like a year+ and just saw your message. First off, I've only had huge negatives with using a 1000hz keyboard. Using a 1000hz polling rate keyboard instead of a 125hz one always seems to just degrade perormance of the mouse for me. It gets a little more technical because some motherboards with USB transaction translators favor having devices all using the same USB specification (2.0+2.0 and not 2.0 + 1.1) plugged into the same hub. But even then, a 1000hz keyboard will probably still be worse overall.
I would not touch 1000hz keyboards at all and use only 125hz ones, or try out a PS/2 keyboard instead if you want to experiment with something different. You tend to get the best mouse movement with having the mouse as the only USB device plugged in on the entire system. The PS/2 port is basically designed from the ground up for keyboards in the first place. Not sure why USB keyboards are even sold beyond the ability to support hot swapping. As for brand of PS/2 keyboard, I tried several and they all generally felt like an improvement over having a USB keyboard plugged in with the EXCEPTION of the Steelseries one.
No idea what's going on with that keyboard, but something is not right about it and it feels even worse than having a 1000hz USB keyboard plugged in. If you want to play it safe, I would attempt to purchase as barebones of PS/2 keyboard as possible and aim specifically for one that only supports 125hz when used in USB mode as well. The coolermaster one for instance has some dumb "turbo mode PS/2 repeat rate" setting and if that setting is set any higher than 1x, your mouse feels bogged down. It's just flooding the ps/2 port with spam blowing up your system or something. Go for as barebones of PS/2 as possible if you want to try and experiment with PS/2.