Z690 usb/timer issue

[Thatweirdinputlag]

Regarding Z690 specifically Asus,

Their high end lineup is having issues with memory training & stability for a while now. Users with boards like the Apex and even the Extreme are experiencing those issues. I purchased an Apex as my first step in upgrading then I returned it back "without checking if it had the same problem or not" just because of how many users have reported insane amount of instability. Mainly something that goes within the realm of "I set my XMP profile, i did stress testing, all was okay, then the next days it just goes unstable". DDR5 is still new and I can only hope that it gets better with time and some bios updates, but you can try a couple of things in bios mentioned in BZ's video "Actually Hardcore Overclocking - Youtube" "https://youtu.be/pRw-RXd93Dc" Mainly within Memory Training Algorithms.

[DPRTMELR]

I've been trying to get bios hpet to work nicely for the 0.5ms timer (which I've achieved previously with a simple setplatformclock no) but this is just cool. I've been to scouring the interwebs to find someone to mod my bios for me(that or spend another $400 on msi board) lol.


Among the ones not mentioned on this thread:

USB PDO Programming | VarStore: PchSetup

I thought this could get me back 2 hubs under 2 controllers properly without virtual nonsense. It doesn't break anything but does enable a bunch of onboard devices that I did not enable, I guess it simply defaults back to asus' "reference" pci lane designs.

XHCI Legacy Support | VarStore: UsbSupport

Disabling this option (this is different from legacy usb support) will keep the usb from loading on the bios (shows 0 xhci loaded D=), so this is probably the ultimate xhci/ehci hand off setting that will make your mouse performance better by 1%. I will try it out them out as the last step after I am done with other bios settings.


-edit

ermmm with HPET disabled on bios, I can't get the windows timer to lock onto 0.5ms and it will generate empty lags that's not from any workloads. This is happening independent to the windows hpet setting or system's RTC timer. Computer is unusable with RTC timer/bios hpet disabled, while acpi timer would make things floaty overall (with the accompanying latency). Not the biggest deal but I would appreciate it someone got a pointer for me on z590 apex.

[Slender]

I've been trying to get bios hpet to work nicely for the 0.5ms timer (which I've achieved previously with a simple setplatformclock no) but this is just cool. I've been to scouring the interwebs to find someone to mod my bios for me(that or spend another $400 on msi board) lol.


Among the ones not mentioned on this thread:

USB PDO Programming | VarStore: PchSetup

I thought this could get me back 2 hubs under 2 controllers properly without virtual nonsense. It doesn't break anything but does enable a bunch of onboard devices that I did not enable, I guess it simply defaults back to asus' "reference" pci lane designs.

XHCI Legacy Support | VarStore: UsbSupport

Disabling this option (this is different from legacy usb support) will keep the usb from loading on the bios (shows 0 xhci loaded D=), so this is probably the ultimate xhci/ehci hand off setting that will make your mouse performance better by 1%. I will try it out them out as the last step after I am done with other bios settings.


-edit

ermmm with HPET disabled on bios, I can't get the windows timer to lock onto 0.5ms and it will generate empty lags that's not from any workloads. This is happening independent to the windows hpet setting or system's RTC timer. Computer is unusable with RTC timer/bios hpet disabled, while acpi timer would make things floaty overall (with the accompanying latency). Not the biggest deal but I would appreciate it someone got a pointer for me on z590 apex.

for my opinion, u can just use grub for change hidden settings.
For long time test im prefer
hpet bios on
useclock/tick no (deletevalue)
0.5ms timer.
And RTC source timer.

[Anonymous27]

Can someone tell me why this is happening?
No settings change, it's just that sometimes everything is fine, but it's very rare, 1 out of 10 reboots, it immediately feels like the gameplay is smooth, according to tests it really is. This happens not only in cs go. This happened on 3 of my different systems with completely different components.

That's strange.. I recently noticed same thing with Valorant.

For me, very often enemies will move really fast and jittery, game will feel a bit heavy and hard to react as everyone else seems to be playing like a ferrari. I play a game on my main immortal 2, top frag as omen and kick ass. Then I play later on smurf and I'm getting whooped by golds.

Sometimes when I have a really nice smooth game, I notice that every game that session is smooth UNTIL I restart the game. Then it's back to shite. I don't know the exact probabilities, but there might be something that's causing game to feel wonky the moment you start it up.

What is it? I don't know. Maybe someone has an idea or a way to test so we can move in the right direction. If this is happening to CS:GO AND Valorant, then it's not game engine or netcode linked, but there has to be another common factor.

[DPRTMELR]

Can someone tell me why this is happening?
No settings change, it's just that sometimes everything is fine, but it's very rare, 1 out of 10 reboots, it immediately feels like the gameplay is smooth, according to tests it really is. This happens not only in cs go. This happened on 3 of my different systems with completely different components.

That's strange.. I recently noticed same thing with Valorant.

I can't speak for z690 but im 99% sure asus's z590 apex's got timers messed up too on the bios somehow. (like a lot of other things =\) I did a quick check after resetting bios hpet and confirmed the 0.5ms timer, then went onto try some other usb ports and pcie things, came back to play some games and the exact things you guys are describing would happen.

sluggish games that barely holds together then just "breaks" after coming back from an alt tab, I was able to get some relief by unsetting core 0 from the game. But allowing hpet on bios is what fixed it for me . With the system hpet turned off latencymon wouldn't show anything above 40 but my computer would still struggle.. I am guessing there's invisible 0.018ms or so that should've been accounted for on your timers but weren't showing up properly and it's screwing things up networks probably? or the clock boosting. It kind of felt like it got worse as time went on but it wasn't the thermals for sure. this all happened after fresh bios defaults reset and summerproof overclocks because last time around my timers showed up wrong but I lagged in a different pattern lol

Other than the misaligned timer, I guess for you two maybe the mouse overclock mod(the one with device filtering) can be a possibility, or something along the lines of async9 misbehaving since the windows update.

sadface me coming to the realization that the misrouted asmedia ports sucks so much ass that I am certain I am getting noises from there and the indiscernable micro pops are hurting my ears and making my night crappy. I've just turned them all off, back to 4k for me on the red intel ports D=

[Thatweirdinputlag]

Can someone tell me why this is happening?
No settings change, it's just that sometimes everything is fine, but it's very rare, 1 out of 10 reboots, it immediately feels like the gameplay is smooth, according to tests it really is. This happens not only in cs go. This happened on 3 of my different systems with completely different components.

That's strange.. I recently noticed same thing with Valorant.

I can't speak for z690 but im 99% sure asus's z590 apex's got timers messed up too on the bios somehow. (like a lot of other things =\) I did a quick check after resetting bios hpet and confirmed the 0.5ms timer, then went onto try some other usb ports and pcie things, came back to play some games and the exact things you guys are describing would happen.

sluggish games that barely holds together then just "breaks" after coming back from an alt tab, I was able to get some relief by unsetting core 0 from the game. But allowing hpet on bios is what fixed it for me . With the system hpet turned off latencymon wouldn't show anything above 40 but my computer would still struggle.. I am guessing there's invisible 0.018ms or so that should've been accounted for on your timers but weren't showing up properly and it's screwing things up networks probably? or the clock boosting. It kind of felt like it got worse as time went on but it wasn't the thermals for sure. this all happened after fresh bios defaults reset and summerproof overclocks because last time around my timers showed up wrong but I lagged in a different pattern lol

Other than the misaligned timer, I guess for you two maybe the mouse overclock mod(the one with device filtering) can be a possibility, or something along the lines of async9 misbehaving since the windows update.

sadface me coming to the realization that the misrouted asmedia ports sucks so much ass that I am certain I am getting noises from there and the indiscernable micro pops are hurting my ears and making my night crappy. I've just turned them all off, back to 4k for me on the red intel ports D=

I can't remember which windows build/version I was using, but I do remember when ppl started disabling platform tick and using resolution timer to force 0.5, I would get 0.496 without the cmd platformtick tweak, then I'd always get flat 0.5 with the cmd tweak. On windows 11 it just doesn't work the same, I get random numbers everytime.

[DPRTMELR]

I can't remember which windows build/version I was using, but I do remember when ppl started disabling platform tick and using resolution timer to force 0.5, I would get 0.496 without the cmd platformtick tweak, then I'd always get flat 0.5 with the cmd tweak. On windows 11 it just doesn't work the same, I get random numbers everytime.

Yeah I am not worried about it. Things were great before I tried bios hpet so it's really whatever.

In fact I just sent in an email to asus regarding my hyperthreaded usb controller and waiting for response. It's probably not gonna go anywhere because nobody really cares and it's easy to call a "ur computers not ready for 8khz mouse please buy a 12900k and our latest z690 motherboard", but I came to the conclusion that I could probably debug this with the available commands but it's gonna come out cheaper for me to just change the board outright.

Their high end lineup is having issues with memory training & stability for a while now. Users with boards like the Apex and even the Extreme are experiencing those issues. I purchased an Apex as my first step in upgrading then I returned it back "without checking if it had the same problem or not" just because of how many users have reported insane amount of instability. Mainly something that goes within the realm of "I set my XMP profile, i did stress testing, all was okay, then the next days it just goes unstable". DDR5 is still new and I can only hope that it gets better with time and some bios updates, but you can try a couple of things in bios mentioned in BZ's video "Actually Hardcore Overclocking - Youtube" "https://youtu.be/pRw-RXd93Dc" Mainly within Memory Training Algorithms.

It's the rtl iol, you put in one value too low manually and you are gonna see some wild stuffs happening, but asus's option would train them to 67/4 or 64/7 based on your dram voltages and that asus ram optimization setting above the 4 big timing is the activator i think, but it would only get applied after a hard shut off and a startup and not reboot (even with fastboot disabled) because you thought that settings safe, but it's really not. Then you would try the other mode just in case but again, nothing would change until a hard reboot, so picnic starts.

[Slender]

Can someone tell me why this is happening?
No settings change, it's just that sometimes everything is fine, but it's very rare, 1 out of 10 reboots, it immediately feels like the gameplay is smooth, according to tests it really is. This happens not only in cs go. This happened on 3 of my different systems with completely different components.

That's strange.. I recently noticed same thing with Valorant.

I can't speak for z690 but im 99% sure asus's z590 apex's got timers messed up too on the bios somehow. (like a lot of other things =\) I did a quick check after resetting bios hpet and confirmed the 0.5ms timer, then went onto try some other usb ports and pcie things, came back to play some games and the exact things you guys are describing would happen.

sluggish games that barely holds together then just "breaks" after coming back from an alt tab, I was able to get some relief by unsetting core 0 from the game. But allowing hpet on bios is what fixed it for me . With the system hpet turned off latencymon wouldn't show anything above 40 but my computer would still struggle.. I am guessing there's invisible 0.018ms or so that should've been accounted for on your timers but weren't showing up properly and it's screwing things up networks probably? or the clock boosting. It kind of felt like it got worse as time went on but it wasn't the thermals for sure. this all happened after fresh bios defaults reset and summerproof overclocks because last time around my timers showed up wrong but I lagged in a different pattern lol

Other than the misaligned timer, I guess for you two maybe the mouse overclock mod(the one with device filtering) can be a possibility, or something along the lines of async9 misbehaving since the windows update.

sadface me coming to the realization that the misrouted asmedia ports sucks so much ass that I am certain I am getting noises from there and the indiscernable micro pops are hurting my ears and making my night crappy. I've just turned them all off, back to 4k for me on the red intel ports D=

I can't remember which windows build/version I was using, but I do remember when ppl started disabling platform tick and using resolution timer to force 0.5, I would get 0.496 without the cmd platformtick tweak, then I'd always get flat 0.5 with the cmd tweak. On windows 11 it just doesn't work the same, I get random numbers everytime.

stop using useplatformtick yes, cause unstable mouse and lower perfomance.
for me, on z690 i take 0.4966 0.4999 0.5
it is good, but see strange.

[Thatweirdinputlag]

I can't remember which windows build/version I was using, but I do remember when ppl started disabling platform tick and using resolution timer to force 0.5, I would get 0.496 without the cmd platformtick tweak, then I'd always get flat 0.5 with the cmd tweak. On windows 11 it just doesn't work the same, I get random numbers everytime.

Yeah I am not worried about it. Things were great before I tried bios hpet so it's really whatever.

In fact I just sent in an email to asus regarding my hyperthreaded usb controller and waiting for response. It's probably not gonna go anywhere because nobody really cares and it's easy to call a "ur computers not ready for 8khz mouse please buy a 12900k and our latest z690 motherboard", but I came to the conclusion that I could probably debug this with the available commands but it's gonna come out cheaper for me to just change the board outright.

Their high end lineup is having issues with memory training & stability for a while now. Users with boards like the Apex and even the Extreme are experiencing those issues. I purchased an Apex as my first step in upgrading then I returned it back "without checking if it had the same problem or not" just because of how many users have reported insane amount of instability. Mainly something that goes within the realm of "I set my XMP profile, i did stress testing, all was okay, then the next days it just goes unstable". DDR5 is still new and I can only hope that it gets better with time and some bios updates, but you can try a couple of things in bios mentioned in BZ's video "Actually Hardcore Overclocking - Youtube" "https://youtu.be/pRw-RXd93Dc" Mainly within Memory Training Algorithms.

It's the rtl iol, you put in one value too low manually and you are gonna see some wild stuffs happening, but asus's option would train them to 67/4 or 64/7 based on your dram voltages and that asus ram optimization setting above the 4 big timing is the activator i think, but it would only get applied after a hard shut off and a startup and not reboot (even with fastboot disabled) because you thought that settings safe, but it's really not. Then you would try the other mode just in case but again, nothing would change until a hard reboot, so picnic starts.

Well, not quite. Asus Mode 1 & 2 are not the culprit of the DDR5 XMP problems on those boards. An entire batch for the Z690 Apex was having insane quality issues, and was mainly because of the signal integrity on one of the DRAM Slots. Imagine a $790 board pushed to the public while one of its main strengths is defected, which is memory overclocking. Sadly Asus did not address this since it did not get enough attention "from High following count accounts" and they decided to sweep it under the rug and offer refurbished boards to users that had the board for more than 2 months or new boards to the people that had it for less! Just imagine you spend close to 3 months trying to figure out why is your DDR5 6400 G.skill ram cannot run at XMP unless you take the frequency down to 5800 or even 5600, all the troubleshooting, buying new ram, time spent, only to get slapped with "Oh hey, you had the motherboard for too long to get a new replacement, we'll send you a refurbished one". That is just insane!
Asus went under fire after its mess and quality control issue on the Z690 Hero, and the burning mosfet. So they probably did not want more of that to come around again on a higher end board.
Source: Asus Rog Forums.

[DPRTMELR]

Well, not quite. Asus Mode 1 & 2 are not the culprit of the DDR5 XMP problems on those boards. An entire batch for the Z690 Apex was having insane quality issues, and was mainly because of the signal integrity on one of the DRAM Slots. Imagine a $790 board pushed to the public while one of its main strengths is defected, which is memory overclocking. Sadly Asus did not address this since it did not get enough attention "from High following count accounts" and they decided to sweep it under the rug and offer refurbished boards to users that had the board for more than 2 months or new boards to the people that had it for less! Just imagine you spend close to 3 months trying to figure out why is your DDR5 6400 G.skill ram cannot run at XMP unless you take the frequency down to 5800 or even 5600, all the troubleshooting, buying new ram, time spent, only to get slapped with "Oh hey, you had the motherboard for too long to get a new replacement, we'll send you a refurbished one". That is just insane!
Asus went under fire after its mess and quality control issue on the Z690 Hero, and the burning mosfet. So they probably did not want more of that to come around again on a higher end board.
Source: Asus Rog Forums.

I didn't even know this was happening D= tbh I thought 12th gen was gonna get tripped by the biglittle play, didn't know ddr5 was going to be this big of a problem. So a physical defect, nice.