Another look at HPET High Precision Event Timer

Discussion in 'Videocards - NVIDIA GeForce Drivers Section' started by Bukkake, Sep 18, 2012.

  1. X7007

    X7007 Ancient Guru

    Messages:
    1,524
    Likes Received:
    16
    GPU:
    Inno3D 1080GTX
    I noticed when changing the DisableDynamicTick Yes changes the mouse movement for the worse , less accurate . Also for some reason the Enhance Pointer precision in the windows effects more when this it's Yes . less noticed when it's No . The mouse pointer is like floating .... now I understand why my mouse behave like this on my Intel I7 3770K , and I didn't connect it to this .
     
  2. X7007

    X7007 Ancient Guru

    Messages:
    1,524
    Likes Received:
    16
    GPU:
    Inno3D 1080GTX
    Can anyone check their Windows timer when they enter a game ? does it 0.9989 or 0.4498 ? I don't understand why when gaming it goes to 0.9989 instead 0.4498 , only when I playback movie it goes to the max..
     
  3. mbk1969

    mbk1969 Ancient Guru

    Messages:
    9,677
    Likes Received:
    6,573
    GPU:
    GF RTX 2070 Super
    I suddenly understood the thing - you was talking not about "MHz" of HPET, but about "ms" of system timer resolution going from 0.5 ms up to 15 ms.
     
  4. mbk1969

    mbk1969 Ancient Guru

    Messages:
    9,677
    Likes Received:
    6,573
    GPU:
    GF RTX 2070 Super
    Most probably game(s) just ordered the resolution of "1ms" and "0.9989" is just result of conversion (from "100ns units" to "ms") without proper floor/ceiling. And media player app ordered maximum "0.5ms" resolution.

    Win API function for setting system timer resolution accepts any input from "0.5" to "15.6".
     

  5. mbk1969

    mbk1969 Ancient Guru

    Messages:
    9,677
    Likes Received:
    6,573
    GPU:
    GF RTX 2070 Super
    I guess that dynamic tick setting is about tickless kernel mode which is native mode since Win8. If that is correct then probably disabling dynamic ticks you switch OS kernel from native to legacy mode.

    https://arstechnica.com/information...-on-the-inside-under-the-hood-of-windows-8/2/
     
  6. hemla

    hemla Member Guru

    Messages:
    171
    Likes Received:
    10
    GPU:
    nvidia
    Yes, it was my bad. On 1803 I am locked at 3.5MHz with timer working at around 1.0ms no matter the mode. But in Windows 1709(or 1703, can't remember) it was changing when I enabled game mode.
     
  7. lamparcicho

    lamparcicho Member

    Messages:
    24
    Likes Received:
    1
    GPU:
    Evga 980Ti sc+
    Hello, on Redstone 5 after clean install I see 10MHz in WinTimerTester and PcClockTiming. On Redstone 4 I have 3.418MHz. No modification in bcdedit on both systems, hpet off in bios. Bug or what? Please guys check this if you have new update and some time. Cheers!
     
  8. Mustang104

    Mustang104 Active Member

    Messages:
    69
    Likes Received:
    10
    GPU:
    nVidia 2080 RTX
    I have noticed this too, stuck at 10mhz no matter what I change...

    [​IMG]
     
    Last edited: Oct 3, 2018
    lamparcicho likes this.
  9. EdKiefer

    EdKiefer Ancient Guru

    Messages:
    2,516
    Likes Received:
    241
    GPU:
    MSI 970 Gaming 4G
    So is this a feature or bug in 1809?

    I am seeing 10mhz with HPET enabled in bios (always had it enabled).
     
    lamparcicho likes this.
  10. mbk1969

    mbk1969 Ancient Guru

    Messages:
    9,677
    Likes Received:
    6,573
    GPU:
    GF RTX 2070 Super
    May be this is due to these things
    https://blogs.technet.microsoft.com/networking/2018/07/18/top10-ws2019-hatime/
    Clock Source Stability
    Our final accuracy-based improvement actually affects the stability of the clock. It’s not enough to have an accurate clock occasionally; you must maintain that accuracy over long periods of time. It’s important to understand that a host system receives time “samples” from its time server, however it does not immediately apply these samples to the clock.

    You can imagine that if a time sample is subject to variable network delay (among other unpredictable network challenges) and we immediately stepped the clock to match every time sample, the clock would likely be incorrect fairly often – it could even move backwards – a problem that would certainly make for a rainy day in the life of an IT Pro…

    Instead we take multiple time samples, eliminate the outliers, and discipline the clock with the goal of bringing the system closer and closer to synchronization with the time server.

    Disciplining the clock entails making adjustments to gradually converge on the correct time. Ultimately there is a natural limit to how small of a change we can make but the key is that smaller is better. Just how granular can we get? This is a complicated question but is based on the frequency of the QPC clock.

    For a more in-depth look at this subject including QPC, please reference this article.
    Previous versions of Windows allowed for a QPC granularity (the smallest change we could make to the system clock) of 6.4 µs/second (microseconds / second). In Windows Server 2019, the QPC granularity drops to 100 nanoseconds / second! This is akin to the difference in clarity between 480p and 4K television. There is much finer granularity in the 4K picture!

    So why does all this matter? Well accuracy as measured over time is reflective of your stability; not only can we hit the bulls-eye, we can hit the bulls-eye over and over again. In a 3.5-day measurement, our partners at Sync-N-Scale measured, and NIST corroborated, Windows Server 2019 pre-release bits. In the picture below, notice the MIN Time Offset reports 41µs (microseconds) RMS diverged from UTC(NIST)!

    [​IMG]

    Note: The AVG method involves comparing the system under test to UTC(NIST) every 10 seconds, then averaging these measurements for 10 minutes (60 readings). UTC(NIST) is available with 0.0001 ms resolution. The difference between the two 10-minute averages is the difference between the time broadcast by the server and UTC(NIST).

    The MIN method involves comparing each NTP server to UTC(NIST) every 10 seconds for a 10 minute interval (60 measurements). However, only one of the 60 measurements is saved, the one with the shortest round trip delay. This method is based on the assumption that NTP measurements with the shortest round trip delays provide the best estimate of the true time difference.
     
    mirh, lamparcicho and Klementh like this.

  11. Mustang104

    Mustang104 Active Member

    Messages:
    69
    Likes Received:
    10
    GPU:
    nVidia 2080 RTX
    So that's it then, we're stuck on 10mhz clock? To be honest it doesn't seem to be affecting performance or anything, just a little concerned that it's locked at this speed.
     
  12. mbk1969

    mbk1969 Ancient Guru

    Messages:
    9,677
    Likes Received:
    6,573
    GPU:
    GF RTX 2070 Super
    Will it stay at 10MHz if you disable HPET in BIOS? Would you please test it for us?
     
  13. Mustang104

    Mustang104 Active Member

    Messages:
    69
    Likes Received:
    10
    GPU:
    nVidia 2080 RTX
    Same dilemma, I don't have the option to enable or disable in BIOS either so I'm at the mercy of the OS.
     
  14. EdKiefer

    EdKiefer Ancient Guru

    Messages:
    2,516
    Likes Received:
    241
    GPU:
    MSI 970 Gaming 4G
    I just tested again, HPET on =10mhz and HPET off in bios = 10mhz to with wintimertester tool.
    So doesn't seem to change anything on my end Intel P8Z77 V Pro i5-3570k.
     
  15. mbk1969

    mbk1969 Ancient Guru

    Messages:
    9,677
    Likes Received:
    6,573
    GPU:
    GF RTX 2070 Super
    If they have not forced HPET then I would not worry about that. All apps which shows that value just calls API function QueryPerformanceFrequency, so in RS5 this function just returns something new (and different).
     

  16. EdKiefer

    EdKiefer Ancient Guru

    Messages:
    2,516
    Likes Received:
    241
    GPU:
    MSI 970 Gaming 4G
    I am not worried :)
    I have always let HPET on in bios (default) and let windows do its thing, which before was like 3.3mhz or something.
     
  17. Blackfyre

    Blackfyre Maha Guru

    Messages:
    1,071
    Likes Received:
    61
    GPU:
    RTX 2070 Super
    So what's the general consensus these days with the latest Windows 10, how does HPET enabled or disabled impact latency & games? I've had it disabled for a long time now and might do a clean install of 1809, should I enable it in the BIOS before formatting? Leave it disabled?
     
  18. mbk1969

    mbk1969 Ancient Guru

    Messages:
    9,677
    Likes Received:
    6,573
    GPU:
    GF RTX 2070 Super
    If HPET is used by Windows as an implementation of QueryPerformanceCounter API function it could introduce increased latencies. So if you are paranoid you can leave HPET disabled in BIOS. If you are not and you trust MS kernel programmers to be professionals you can enable HPET in BIOS believing Windows to use it only if there is no better timer implementation.
     
    Smough and AsiJu like this.
  19. mbk1969

    mbk1969 Ancient Guru

    Messages:
    9,677
    Likes Received:
    6,573
    GPU:
    GF RTX 2070 Super
    I have the setting to switch HPET "On/Off" in BIOS, and I prefer to set it "On" in BIOS. The thought is: any program can use HPET directly, and Windows should not use it by default.

    Theoreticlly speaking you can forbid Windows to use any so called platform timers with "bcdedit /set useplatformclock no"
     
  20. lamparcicho

    lamparcicho Member

    Messages:
    24
    Likes Received:
    1
    GPU:
    Evga 980Ti sc+
    Hello guys :) What advantage or disadvantage gaves us "bcdedit /set tscsyncpolicy Enhanced"? Where and how I can see or I can feel difference? I mean better responsiveness system or faster system, maybe lowest lag or what? I am very curious.

    Cheers!
     

Share This Page