revisiting hpet bcdedit tweaks: what are your timer bench results and settings?

[zoulztealer]

Or Windows builds.

Surprising Observation! HPET On vs. Off - Page 3 - Blur Busters Forums

Using above settings X99

HPET can be used as a timer and/or counter. Usually RTC is used as a timer if HPET is disabled (BIOS or register) and platformtick is requested. It's unusual that it runs at a timer resolution of ~0.4882ms and presumably Windows counts the number of interrupts before doing something other than just acknowledging, so 32 interrupts for 15.6ms for example. This also means min and max resolution ca be set lower or higher than OS limits.

Example at 125ms

Also no noticeable pops, cracks or whistles with audio with that setting although testing was very limited.

Seems disabledynamictick on this system causes timer interrupts to only only occur on CPU 0 instead of any CPU when LAPIC is used.

very interesting!

yes, dpc lat can be quite high without issues depending on the settings.

what are those tools you are using in the screens (ntresolution and win reported clocks nt) and can they be downloaded somewhere or are they already supplied within windows?

 

[:D:D]

ntresolution just uses the windows calls NtQueryTimerResolution and NtSetTimerResolution, Win registered clocks is something I wrote years ago but never finished and as such isn't a universal user tool. I'm only a hobby programmer (self taught) so it would need a lot of work from me, not worth the effort. MS does impress with their clock frequency calculations.

Edit: Forgot to mention, the synthetic timers are those produced by Hyper-V while HPET is a real timer.

 

[zoulztealer]

ntresolution just uses the windows calls NtQueryTimerResolution and NtSetTimerResolution, Win registered clocks is something I wrote years ago but never finished and as such isn't a universal user tool. I'm only a hobby programmer (self taught) so it would need a lot of work from me, not worth the effort. MS does impress with their clock frequency calculations.

Edit: Forgot to mention, the synthetic timers are those produced by Hyper-V while HPET is a real timer.

woah, thats cool. you should definitely finish them and can i have a copy win reg clocks?

do you know which timers exactly are considered as synthetic and which are not?

and whats the impact of hyperv on timers and how to completely disable hyperv?

i can see 3 timers in ms interrupt policy tool. they all have spread across all processors policy. i wonder if one timer needs to run on a specific cpu or if they could be spread over all processors (all processors in machine):

 

[izy]

Before a reboot it was only yellow even on idle now its constant @ 500 with chrome tabs opened.

With WoW opened in background:

After i enabled MSI for the GPU, game opened in the background (for some reason nvidia driver sets this to linebased)

Also TimeBench does something funny

 

[zoulztealer]

found some things new but also some "errors" in there

 

[:D:D]

do you know which timers exactly are considered as synthetic and which are not?

Ones that are virtual/artificial, not real. Real ones are accessed from HW. There's a bunch of reserved MSR's from 0x40000000 that are used by MS for this purpose and do not exist on the CPU.
Maybe this paper will help.

https://raw.githubusercontent.com/Microsoft/Virtualization-Documentation/master/tlfs/Hypervisor%20Top%20Level%20Functional%20Specification%20v5.0.pdf

and whats the impact of hyperv on timers and how to completely disable hyperv?

I like to have access to my HW, some of that access is disabled by Hyper-V which is why I disabled it. Since that also means no synthetic timers I have no experience of their impact.

Disabling was a little more involved than just using BCD. There is a MS PS script to help. Disabled but still lurking IIRC. Don't know if you need to go that far for disabling synthetic timers though.

Download Device Guard and Credential Guard hardware readiness tool from Official Microsoft Download Center

Use this tool to see if your hardware is ready for Device Guard and Credential Guard. You can also use this to enable Device Guard or Credential Guard.

www.microsoft.com

Setting affinity may be a case by case basis and needing testing for different scenarios.

@izy your first run shows timer resolution was set to 0.5ms instead of 1ms and will likely affect results. Unfortunately that value isn't updated in real time so you might not know if some other app has requested a lower time during testing. Looks like TimerBench hasn't been updated in a long time too.

 

[Apocalypsee]

found some things new but also some "errors" in there

The second video is a bit extreme on some part. I did some of the tweaks but now dpclatency is either red or yellow again. And this comes out. Idk what to do anymore. Seems like I did something right then other things pops out

 

[zoulztealer]

The second video is a bit extreme on some part. I did some of the tweaks but now dpclatency is either red or yellow again. And this comes out. Idk what to do anymore. Seems like I did something right then other things pops out

and what exactly did you change? but whatever it was you should be able to reverse it. if you dont have nothing important on that win i wouldnt waste time and just make a fresh install and start again with what you know so far and dont make too many changes at once and keep track of each you apply.

i also wanted to mention daddy madus debloat and optimization script, but beware and do NOT simply run the whole script, but read the code and get out from it what you want:

guide:

GitHub - DaddyMadu/Windows10GamingFocus: This is the Ultimate Windows 10/11 Debloat and Gaming Tweaks Optimization based on ChrisTitusTechScript which is based on multiple debloat scripts and gists from github.

This is the Ultimate Windows 10/11 Debloat and Gaming Tweaks Optimization based on ChrisTitusTechScript which is based on multiple debloat scripts and gists from github. - GitHub - DaddyMadu/Windo...

github.com

script code:

Windows10GamingFocus/win10debloatandgamingtweaks.ps1 at master · DaddyMadu/Windows10GamingFocus

This is the Ultimate Windows 10/11 Debloat and Gaming Tweaks Optimization based on ChrisTitusTechScript which is based on multiple debloat scripts and gists from github. - DaddyMadu/Windows10Gamin...

github.com

as well as atlas.os i havent tried myself yet but reports of it are interesting:

An optimized modification of Windows, designed for enthusiasts.

We designed Atlas for enthusiasts & gamers, improving privacy, usability, and performance. Have the best experience compared to stock, slow Windows.

atlasos.net

heres another tool can do some disabling, also hyperv, not sure how it does it exactly, but there are several different methods apparently with different effects:

Windows 10 Tweak Tool

Download Windows 10 Tweak Tool for free. Disable all bullshit in Windows 10 in a couple of clicks. Windows 10 has a lot of useless services and processes that load the PC, so why not disable them. With Windows 10 Tweak Tool, this can be done in a couple of clicks.

sourceforge.net

How to Uninstall or Disable Hyper-V in Windows

Learn how to uninstall Hyper-V in Windows 10 and Windows Server 2019 by using the graphical user interface or the command line.

www.nakivo.com

if you enter systeminfo some nice stuff comes up also showing info about hyperv, i have it uninstalled, but its still enabled somehow:

 

[Apocalypsee]

and what exactly did you change? but whatever it was you should be able to reverse it. if you dont have nothing important on that win i wouldnt waste time and just make a fresh install and start again with what you know so far and dont make too many changes at once and keep track of each you apply.

i also wanted to mention daddy madus debloat and optimization script, but beware and do NOT simply run the whole script, but read the code and get out from it what you want:

guide:

GitHub - DaddyMadu/Windows10GamingFocus: This is the Ultimate Windows 10/11 Debloat and Gaming Tweaks Optimization based on ChrisTitusTechScript which is based on multiple debloat scripts and gists from github.

This is the Ultimate Windows 10/11 Debloat and Gaming Tweaks Optimization based on ChrisTitusTechScript which is based on multiple debloat scripts and gists from github. - GitHub - DaddyMadu/Windo...

github.com

script code:

Windows10GamingFocus/win10debloatandgamingtweaks.ps1 at master · DaddyMadu/Windows10GamingFocus

This is the Ultimate Windows 10/11 Debloat and Gaming Tweaks Optimization based on ChrisTitusTechScript which is based on multiple debloat scripts and gists from github. - DaddyMadu/Windows10Gamin...

github.com

as well as atlas.os i havent tried myself yet but reports of it are interesting:

An optimized modification of Windows, designed for enthusiasts.

We designed Atlas for enthusiasts & gamers, improving privacy, usability, and performance. Have the best experience compared to stock, slow Windows.

atlasos.net

heres another tool can do some disabling, also hyperv, not sure how it does it exactly, but there are several different methods apparently with different effects:

Windows 10 Tweak Tool

Download Windows 10 Tweak Tool for free. Disable all bullshit in Windows 10 in a couple of clicks. Windows 10 has a lot of useless services and processes that load the PC, so why not disable them. With Windows 10 Tweak Tool, this can be done in a couple of clicks.

sourceforge.net

How to Uninstall or Disable Hyper-V in Windows

Learn how to uninstall Hyper-V in Windows 10 and Windows Server 2019 by using the graphical user interface or the command line.

www.nakivo.com

if you enter systeminfo some nice stuff comes up also showing info about hyperv, i have it uninstalled, but its still enabled somehow:

View attachment 363028

I didn't do much for the time being because the lack of time on my side. I only go as far as disabling several devices in Device Manager I don't think that affect anything related to that. The spikes are minor only a bit higher than 2000us as to off the chart previously so I didn't notice any audio stutter or visual glitches while gaming or watching video. That nvidia perhaps because I was browsing and kept switching between iGPU and dGPU.

EDIT: It seems OK today but it reverts back to it's old problems, ACPI.sys.

 

[zoulztealer]

I didn't do much for the time being because the lack of time on my side. I only go as far as disabling several devices in Device Manager I don't think that affect anything related to that. The spikes are minor only a bit higher than 2000us as to off the chart previously so I didn't notice any audio stutter or visual glitches while gaming or watching video. That nvidia perhaps because I was browsing and kept switching between iGPU and dGPU.

EDIT: It seems OK today but it reverts back to it's old problems, ACPI.sys.

the acpi.sys is a great issue of many laptops. you have a dell laptop too or which brand / model exactly?

go to powershell admin and then launch:

1. net start (lists all active services)

2. tasklist (lists all active processes)

and post your results here.

i believe for your system it probably would be best if you enable hpet in bios but disallow it in windows with either bcdedit or timerbench or both. you said you enabled hpet first and it got worse but probably it wasnt defined probably in bios, bcdedit and timerbench / os. you will probably have to use one of the combos mentioned before.

i suggest you either try combo 3 or 4 - or 2.

1. TSC + TSC without desync: bcdedit /deletevalue useplatformclock - bcdedit /deletevalue useplatformtick - make sure HPET is enabled in BIOS

2. TSC + RTC: bcdedit /deletevalue useplatformclock - bcdedit /set useplatformtick Yes

3. HPET + RTC: bcdedit /set useplatformclock Yes - bcdedit /set useplatformtick Yes - make sure HPET is enabled in BIOS

4. PMT + RTC: bcdedit /set useplatformclock Yes - bcdedit /set useplatformtick Yes - make sure HPET is disabled in BIOS

of course bios update and fresh windows install etc could also help you but as long as you dont have any issues anymore there is no need to do anything about it and you can just leave everything as it is if you can play games, audio and video at the same time without dropouts.

id also recommend you always have islc on running in the background with custom timer res 0.5

 

[Apocalypsee]

the acpi.sys is a great issue of many laptops. you have a dell laptop too or which brand / model exactly?

go to powershell admin and then launch:

1. net start (lists all active services)

2. tasklist (lists all active processes)

and post your results here.

i believe for your system it probably would be best if you enable hpet in bios but disallow it in windows with either bcdedit or timerbench or both. you said you enabled hpet first and it got worse but probably it wasnt defined probably in bios, bcdedit and timerbench / os. you will probably have to use one of the combos mentioned before.

i suggest you either try combo 3 or 4 - or 2.

1. TSC + TSC without desync: bcdedit /deletevalue useplatformclock - bcdedit /deletevalue useplatformtick - make sure HPET is enabled in BIOS

2. TSC + RTC: bcdedit /deletevalue useplatformclock - bcdedit /set useplatformtick Yes

3. HPET + RTC: bcdedit /set useplatformclock Yes - bcdedit /set useplatformtick Yes - make sure HPET is enabled in BIOS

4. PMT + RTC: bcdedit /set useplatformclock Yes - bcdedit /set useplatformtick Yes - make sure HPET is disabled in BIOS

of course bios update and fresh windows install etc could also help you but as long as you dont have any issues anymore there is no need to do anything about it and you can just leave everything as it is if you can play games, audio and video at the same time without dropouts.

id also recommend you always have islc on running in the background with custom timer res 0.5

My laptop is ASUS TUF A15, to be specific ASUS FX506HCB i5-11400H, 16GB DDR4-3200, 4GB RTX 3050.

I read in reddit there is something about new intel storage controller, VMD that causing it. If I disable this (which I already tried) I need to reinstall windows. As much as I like to reinstall Windows I really don't have much time to do it. I need to think which version to install first. And I read this fix isn't 100% as there are ppl already reinstall windows and still having this issue.

There isn't any HPET setting in BIOS, even on unlocked BIOS which all setting is shown. I don't think I'll be doing HPET thing again after what I did previously which it causes massive stutter and frame pacing issue. Custom timer is already at 0.5 sec. I don't believe it's HPET issue to begin with, this is something else.

From what I been testing there isn't any dropout from my limited testing yesterday, but I will do further testing. This dropout is totally random and weirdly the DPC problem shifted quite rapidly in-between system restarts.

 

[zoulztealer]

My laptop is ASUS TUF A15, to be specific ASUS FX506HCB i5-11400H, 16GB DDR4-3200, 4GB RTX 3050.

I read in reddit there is something about new intel storage controller, VMD that causing it. If I disable this (which I already tried) I need to reinstall windows. As much as I like to reinstall Windows I really don't have much time to do it. I need to think which version to install first. And I read this fix isn't 100% as there are ppl already reinstall windows and still having this issue.

There isn't any HPET setting in BIOS, even on unlocked BIOS which all setting is shown. I don't think I'll be doing HPET thing again after what I did previously which it causes massive stutter and frame pacing issue. Custom timer is already at 0.5 sec. I don't believe it's HPET issue to begin with, this is something else.

From what I been testing there isn't any dropout from my limited testing yesterday, but I will do further testing. This dropout is totally random and weirdly the DPC problem shifted quite rapidly in-between system restarts.

if you want to test for dpc issues, try playing a game while ripping + encoding audio / video with makemkv or something in the background. if you have can still play the game normally everything is fine.

I like to have access to my HW, some of that access is disabled by Hyper-V which is why I disabled it. Since that also means no synthetic timers I have no experience of their impact.

thats why i "imagined" to force platform clock and tick on so that all timers come from hw and not sw / os / synthetic etc.


my mind is blown from these videos:

more on his channel:

Savitarax

My channel is focused on education and information about computers. I like to teach people how to get the most out of their PC and I like to make sure to always have a good credible history of information that I am referencing. I value operformance and efficiency highly and I seek to let people...

www.youtube.com

 

[ShrimpBrime]

if you want to test for dpc issues, try playing a game while ripping + encoding audio / video with makemkv or something in the background. if you have can still play the game normally everything is fine.


thats why i "imagined" to force platform clock and tick on so that all timers come from hw and not sw / os / synthetic etc.


my mind is blown from these videos:

more on his channel:

Savitarax

My channel is focused on education and information about computers. I like to teach people how to get the most out of their PC and I like to make sure to always have a good credible history of information that I am referencing. I value operformance and efficiency highly and I seek to let people...

www.youtube.com

Interesting. I noticed that the Power Settings Explorer app is exactly what Windows Power Plan (Advanced Options) is. But the application is definitely kick butt. Thanks for sharing

 

[:D:D]

if you enter systeminfo some nice stuff comes up also showing info about hyperv, i have it uninstalled, but its still enabled somehow:

I think it's just showing the requirements are met, not necessarily that it's enabled.

On my system HPET takes precedence so if it's enabled (BIOS) then RTC cannot be selected for timer and PMT cannot be selected for counter.

Thanks for the timer resolution video, now it's clear what was meant by not able to adjust the timer resolution. It still can be adjusted but apps will be ignored for x number of triggers. Similar to how RTC timer (tick) works. Not sure why sleep is being used as it's not meant to be used for accurate timing. When setting a resolution of 0.5ms he gets 0.4992. 2x0.4992 is less than 1ms so 3x0.4992 is used =1.4976ms. Requesting 0.507 sees 0.5068ms set which can satisfy 1ms minimum with 2 triggers. 2x0.5068 gives 1.0136ms.

MS Timer resolution is a lot more aggressive these days and seem to like changing how things work.

 

[zoulztealer]

I think it's just showing the requirements are met, not necessarily that it's enabled.

On my system HPET takes precedence so if it's enabled (BIOS) then RTC cannot be selected for timer and PMT cannot be selected for counter.

Thanks for the timer resolution video, now it's clear what was meant by not able to adjust the timer resolution. It still can be adjusted but apps will be ignored for x number of triggers. Similar to how RTC timer (tick) works. Not sure why sleep is being used as it's not meant to be used for accurate timing. When setting a resolution of 0.5ms he gets 0.4992. 2x0.4992 is less than 1ms so 3x0.4992 is used =1.4976ms. Requesting 0.507 sees 0.5068ms set which can satisfy 1ms minimum with 2 triggers. 2x0.5068 gives 1.0136ms.

MS Timer resolution is a lot more aggressive these days and seem to like changing how things work.

yes, it shows requirements (are met), ive enabled all options for hyperthreading, virtualization, etc in bios but ill do further cleaning on hyperv in the os until its completely purged. vmware even recommends to disable / uninstall hyperv before setting up a virtual machine. hyperv seems to be solely needed for purely ms vms / servers and pretty much pointless for gamers and similar; its not needed generally for vms.

ive noticed different machines report different values for the timers, sometimes they read out smooth numbers like 0.5, sometimes theyre odd like 0.499 or sth. ive seen this change on my machine too from odd to smooth number after i disabled dynamic tick i believe or other bcd options.

i believe these differences may cause synchronization issues over time or on read out by a software depending on the what method, timer, function etc is used and what is defined from the os.

so even i previously had uninstalled hyperv from windows features already it seemed to still be resident in the system. i used these two commands and they were obviously busy doing something and succesfully completed the operation without errors:

Disable-WindowsOptionalFeature -Online -FeatureName Microsoft-Hyper-V-All

DISM /Online /Disable-Feature:Microsoft-Hyper-V

further adding this bcd:

bcdedit /set hypervisorlaunchtype off

My laptop is ASUS TUF A15, to be specific ASUS FX506HCB i5-11400H, 16GB DDR4-3200, 4GB RTX 3050.

I read in reddit there is something about new intel storage controller, VMD that causing it. If I disable this (which I already tried) I need to reinstall windows. As much as I like to reinstall Windows I really don't have much time to do it. I need to think which version to install first. And I read this fix isn't 100% as there are ppl already reinstall windows and still having this issue.

There isn't any HPET setting in BIOS, even on unlocked BIOS which all setting is shown. I don't think I'll be doing HPET thing again after what I did previously which it causes massive stutter and frame pacing issue. Custom timer is already at 0.5 sec. I don't believe it's HPET issue to begin with, this is something else.

From what I been testing there isn't any dropout from my limited testing yesterday, but I will do further testing. This dropout is totally random and weirdly the DPC problem shifted quite rapidly in-between system restarts.

sweet laptop. some devices can be disabled indeed from the device manager and this will help with performance but certain items shouldnt be disabled in there because they may instantly freeze and destroy / corrupt your windows installation or cause other issues with assignment of system ressources and even permanent damage to the mainboard.

you should experiment with one of the combos from melody. as i said it may be your system needs a certain configuration like hpet on in the background but disallowed by os. if you have no option in bios the setting is probably forced always on by default. if you ever reinstall windows be sure you install one of the pre-tweaked versions.

maybe this is something for you too (disables several acpi files):

 

[Apocalypsee]

I did a few things, from ULTIMATE LATENCY GUIDE video I did the ACPI thing, use dynamicticks as per below, did the GPUAffinity and InterruptAffinity thing.

bcdedit /set useplatformclock no
bcdedit /set useplatformtick yes
bcdedit /set disabledynamictick yes

Then I moved on to 2nd video, the TIMER RESOLUTION HAS CHANGED (UPDATED) and the measuresleep timer is...surprising

 

[zoulztealer]

I did a few things, from ULTIMATE LATENCY GUIDE video I did the ACPI thing, use dynamicticks as per below, did the GPUAffinity and InterruptAffinity thing.

bcdedit /set useplatformclock no
bcdedit /set useplatformtick yes
bcdedit /set disabledynamictick yes

Then I moved on to 2nd video, the TIMER RESOLUTION HAS CHANGED (UPDATED) and the measuresleep timer is...surprising

did you get dpclat or latencymon improvements by this?

i dont understand measuresleep exactly but i think your delta is way too high.

checkout this video in the middle where hes benchmarking @4:50 and analyzing @6:50:

youll need to adjust / find the best timer value that results in the lowest delta.

i dont recommend running prime95 like he did though or any from the files he provides.

 

[Apocalypsee]

did you get dpclat or latencymon improvements by this?

i dont understand measuresleep exactly but i think your delta is way too high.

checkout this video in the middle where hes benchmarking @4:50 and analyzing @6:50:

youll need to adjust / find the best timer value that results in the lowest delta.

i dont recommend running prime95 like he did though or any from the files he provides.

DPClat is back to all yellow now with random spikes from either ACPI or w01000. But LatencyMon gives the most improvements. Now the Highest reported ISR routine (the third one) is zero if I just left it idle. Other process latency now drop furter to, usually below 300uS. It's only when I open up few things when the two spikes shows up at random.

I WONT be stress testing as my laptop got crap cooler, it idles at 50C and that is with undervolting

EDIT: Look at this. Simply amazing if it just idle. Never seen anything that flat. Only GoG and Steam in the background

This is when watching video (K-lite codec) while browser in background (Firefox)

There is ACPI.sys spikes when playing games Fallout London but I don't noticed any stutter or audio drop at all. I still don't know what trigger that ACPI.sys though.

 

[zoulztealer]

DPClat is back to all yellow now with random spikes from either ACPI or w01000. But LatencyMon gives the most improvements. Now the Highest reported ISR routine (the third one) is zero if I just left it idle. Other process latency now drop furter to, usually below 300uS. It's only when I open up few things when the two spikes shows up at random.

I WONT be stress testing as my laptop got crap cooler, it idles at 50C and that is with undervolting

EDIT: Look at this. Simply amazing if it just idle. Never seen anything that flat. Only GoG and Steam in the background

This is when watching video (K-lite codec) while browser in background (Firefox)

There is ACPI.sys spikes when playing games Fallout London but I don't noticed any stutter or audio drop at all. I still don't know what trigger that ACPI.sys though.

very nice. seems like you applied the proper bcd combo and other effective tweaking as well.

 

[Apocalypsee]

very nice. seems like you applied the proper bcd combo and other effective tweaking as well.

Thank you for giving me pointers, you been a great help This shows that Windows isn't perfect and plenty of tweaks to make it work better.

I just find out that when watching YouTube there's audio stutter but it's very faint as compared to before, out of all other things....I need to check the measuresleep timer a bit more closely. Timer on CPU-Z reported very close in between the three clocks but htere is still disparity sometimes but it's about 0.01 sec, I just need to figure it out how to sync them, and whether that will fix the dreaded ACPI.sys that plague my laptop for some time now.

EDIT: I monitor the 'audio drop' in YouTube there isn't any DPC latency issue now so idk whats going on right now lol

 

[izy]

Left LM running in the background while doing my stuff , playing a game + working on things, so many alt+tabs or using second monitor and this is what i got , is it normal or something works bad here? On idle everything is fine.

 

[zoulztealer]

Left LM running in the background while doing my stuff , playing a game + working on things, so many alt+tabs or using second monitor and this is what i got , is it normal or something works bad here? On idle everything is fine.

yes it seems all normal, looks same for me. high values are also caused from first time / uncached disk read. latencymon will only give proper values when as much apps drivers services etc are shutdown and have been loaded / cached. there shouldnt be any disk activity any more or drivers etc loaded first time so that latencymon can report proper values. for example when the system is idling after boot and everything has been loaded up and there is almost no disk activity anymore.

Thank you for giving me pointers, you been a great help This shows that Windows isn't perfect and plenty of tweaks to make it work better.

I just find out that when watching YouTube there's audio stutter but it's very faint as compared to before, out of all other things....I need to check the measuresleep timer a bit more closely. Timer on CPU-Z reported very close in between the three clocks but htere is still disparity sometimes but it's about 0.01 sec, I just need to figure it out how to sync them, and whether that will fix the dreaded ACPI.sys that plague my laptop for some time now.

EDIT: I monitor the 'audio drop' in YouTube there isn't any DPC latency issue now so idk whats going on right now lol

thank you, too! thank you everyone who contributed! if you dont need it specifically you could uninstall the klite code as it is known that it can quite cause some issues. this could also be the issue with your browser playback causing audio stutter. it could also be the audio or chipset driver or other things. but as i said before i wanted to analyze the list of your running processes and services because i believe there is a service an app or a driver on your system that is draining your cpu power / causing high cpu % background.

which browser are you using? i recommend operagx or for emergency you can check with epicprivacybrowser it has a very nice built in vpn, antiscript, etc and both are very optimized, has integrated adblocker, antiminer, etc and you can control how much ram and cpu it will use. mine runs with 19% cpu and i have like 200-300 tabs open and i can play games, watch videos on youtube and play dvd at the same etc perfect with no performance issues, two monitors. for media playback you can use vlc media player and you wont need any additional codec pack.

use autoruns64 to disable everything you dont need:

Autoruns - Sysinternals

See what programs are configured to startup automatically when your system boots and you login.

learn.microsoft.com

https://download.sysinternals.com/files/Autoruns.zip

 

[:D:D]

yes, it shows requirements (are met), ive enabled all options for hyperthreading, virtualization, etc in bios but ill do further cleaning on hyperv in the os until its completely purged. vmware even recommends to disable / uninstall hyperv before setting up a virtual machine. hyperv seems to be solely needed for purely ms vms / servers and pretty much pointless for gamers and similar; its not needed generally for vms.

Similar if wanting to use Intel HAXM. The MS Hyper Visor also brings security measures.

ive noticed different machines report different values for the timers, sometimes they read out smooth numbers like 0.5, sometimes theyre odd like 0.499 or sth. ive seen this change on my machine too from odd to smooth number after i disabled dynamic tick i believe or other bcd options.

i believe these differences may cause synchronization issues over time or on read out by a software depending on the what method, timer, function etc is used and what is defined from the os.

Depends on frequency of timer which count only integers (whole numbers). Multiply frequency by 0.0005 and if it isn't a whole number then it will not be the whole number you are looking for. Note that resolution reading is truncated so might not show the whole picture.

so even i previously had uninstalled hyperv from windows features already it seemed to still be resident in the system.

Can check with cpuid as shown in previous posted PDF.

Then I moved on to 2nd video, the TIMER RESOLUTION HAS CHANGED (UPDATED) and the measuresleep timer is...surprising

Windows changed behavior from W10 2004. I think can still adust globally but not using the previous tools. Note, when I tried on W10 22H2 I was still getting some 1.5 times (0.5 delta). IINM the guy said it increased benchmark / FPS when tuned. So this or perfect 0.5000? Unfortunately cannot have your cake and eat it too.

 

[:D:D]

Well I zipped up the app (3KB) to Virus Total and got 13 false positives so maybe best left for my use only.
Top is current resolution, second line min : max delta, 3rd line actual ongoing sleep(1) measurements and 4th line adjustable requested resolution.

You can still check measuresleep app with DPCLat open and should see delta drop from 14 to 1 to show wether global setting is possible. Unfortunately it's fixed around 1ms.

 

[zoulztealer]

Windows changed behavior from W10 2004. I think can still adust globally but not using the previous tools. Note, when I tried on W10 22H2 I was still getting some 1.5 times (0.5 delta). IINM the guy said it increased benchmark / FPS when tuned. So this or perfect 0.5000? Unfortunately cannot have your cake and eat it too.

would there be any issues from forcing globaltimerrequests on in registry? islc says this feature is only for win 11.

You can still check measuresleep app with DPCLat open and should see delta drop from 14 to 1 to show wether global setting is possible. Unfortunately it's fixed around 1ms.

that works for me. so globaltimerrequests seems to work on win 10, too. i get best values around 0.5070 and theres definitely delta changing going on above and below that. as soon as i close dpclat "slept" goes back to 15 again.

(that is with browser and other stuff open)
Resolution: 0.5070ms, Sleep(1) slept 1.0124ms (delta: 0.0124)
Resolution: 0.5070ms, Sleep(1) slept 1.0105ms (delta: 0.0105)
Resolution: 0.5070ms, Sleep(1) slept 1.0129ms (delta: 0.0129)
Resolution: 0.5070ms, Sleep(1) slept 1.0110ms (delta: 0.0110)
Resolution: 0.5070ms, Sleep(1) slept 1.0115ms (delta: 0.0115)
Resolution: 0.5070ms, Sleep(1) slept 1.0119ms (delta: 0.0119)
Resolution: 0.5070ms, Sleep(1) slept 1.0121ms (delta: 0.0121)
Resolution: 0.5070ms, Sleep(1) slept 1.0126ms (delta: 0.0126)
Resolution: 0.5070ms, Sleep(1) slept 1.0130ms (delta: 0.0130)
Resolution: 0.5070ms, Sleep(1) slept 1.0129ms (delta: 0.0129)
Resolution: 0.5070ms, Sleep(1) slept 1.0096ms (delta: 0.0096)
Resolution: 0.5070ms, Sleep(1) slept 1.0119ms (delta: 0.0119)
Resolution: 0.5070ms, Sleep(1) slept 1.0116ms (delta: 0.0116)
Resolution: 0.5070ms, Sleep(1) slept 1.0130ms (delta: 0.0130)
Resolution: 0.5070ms, Sleep(1) slept 1.0120ms (delta: 0.0120)

now when i launch war thunder the timer gets locked to 0.5 res and the islc timer setting gets ignored, so war thunder forces another resolution value over the islc setting. when i shut everything (wt, islc, processlasso) down i get 1.0 res

Resolution: 0.5000ms, Sleep(1) slept 1.5144ms (delta: 0.5144)
Resolution: 0.5000ms, Sleep(1) slept 1.0114ms (delta: 0.0114)
Resolution: 0.5000ms, Sleep(1) slept 1.0102ms (delta: 0.0102)
Resolution: 0.5070ms, Sleep(1) slept 1.0110ms (delta: 0.0110)
Resolution: 0.5070ms, Sleep(1) slept 1.0116ms (delta: 0.0116)
Resolution: 0.5070ms, Sleep(1) slept 15.2065ms (delta: 14.2065)
Resolution: 0.5070ms, Sleep(1) slept 15.1743ms (delta: 14.1743)