Another look at HPET High Precision Event Timer

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

  1. Prophet

    Prophet Master Guru

    Messages:
    822
    Likes Received:
    14
    GPU:
    Msi 680 Gtx Twin Frozr
    From Dpc latency checkers homepage: Windows 8 Compatibility: The DPC latency utility runs on Windows 8 but does not show correct values. The output suggests that the Windows 8 kernel performs badly and introduces a constant latency of one millisecond which is not the case in practice. DPCs in the Windows 8 kernel behave identical to Windows 7. The utility produces incorrect results because the implementation of kernel timers has changed in Windows 8 which causes a side effect with the measuring algorithm used by the utility. Thesycon is working on a new version of the DPC latency utility and will make it available on this site as soon as it is finished.
     
  2. snorge

    snorge Member Guru

    Messages:
    122
    Likes Received:
    0
    GPU:
    gtx 970 SLI
    If it performs identical why is my system unusable with HTEP off in Win 8 but fine in Win 7?
     
  3. HonoredShadow

    HonoredShadow Ancient Guru

    Messages:
    4,284
    Likes Received:
    14
    GPU:
    Gigabyte 1080 ti
    So would I be right in saying that if you have HPET on in bios but not in Windows, that would be the same as disabling HPET?

    Anyway. My results. I left my computer to stand for 5 minutes with nothing open to let the system settle.

    HPET ON in BIOS only:
    [​IMG]

    HPET ON in BIOS and Windows 7:
    [​IMG]

    Now can anyone tell me which would be better out of those two as I really can't tell! HELP! :)
     
    Last edited: Sep 28, 2012
  4. Cyberdyne

    Cyberdyne Ancient Guru

    Messages:
    3,582
    Likes Received:
    302
    GPU:
    2080 Ti FTW3 Ultra
    Both are extremely similar. Try to run it for longer, run more software (watch some youtube, play a game) then see it the better option is more clear. Otherwise, if the result is the same, the differences in results you have are well within margin of error so it does not really matter.
     
    Last edited: Sep 28, 2012

  5. HonoredShadow

    HonoredShadow Ancient Guru

    Messages:
    4,284
    Likes Received:
    14
    GPU:
    Gigabyte 1080 ti
    Ok. Will do. Thanks.
     
  6. Varna

    Varna Active Member

    Messages:
    92
    Likes Received:
    0
    GPU:
    EVGA 690GTX
    DPC is really... inconsistent.

    Jumping around from 240~377. That seems way higher then most people in here though. What are some tips for lowering this?

    Currently have HPET disabled from both the bios and windows. Enabled there is a slight gain in DPC 50~100.
     
  7. HonoredShadow

    HonoredShadow Ancient Guru

    Messages:
    4,284
    Likes Received:
    14
    GPU:
    Gigabyte 1080 ti
    I would say thats all normal tbf
     
  8. MrBonk

    MrBonk Ancient Guru

    Messages:
    3,265
    Likes Received:
    207
    GPU:
    MSI RTX 2080
    I've run into some somewhat interesting results with games since I turned HPET ON/Core Parking Off, restabalized my OC to 4Ghz and upgraded my drivers to 306.23.

    The Last Remnant Benchmark 720p
    305.53 HPET OFF Core Parking On CPU@4.2Ghz
    306.23 HPET ON Core Parking On CPU@4Ghz
    306.23 HPET ON Core Parking Off CPU@4Ghz

    This game is Known to have wildly fluctuating FPS because of developer mistakes, with HPET off and 305.53 you could feel the microstuttering when the FPS would drastically change. With 306.23 and HPET on the changes felt a lot smoother. Though you could tell the FPS was slowing down,it was no longer characterized with a microstutter

    Street Fighter IV Benchmark 720p
    305.53 HPET OFF CP-On 4.2Ghz
    306.23 HPET ON CP-On 4Ghz
    306.23 HPET ON Cp-Off 4Ghz

    Felt no stuttering on any setup. FPS AVG increase is interesting. I want to say within margin of error. but not sure.

    Lost Planet 2 Benchmark 1280x1024
    305.53 HPET-Off CP-On CPU 4.2Ghz Test A
    306.23 HPET-ON CP-On CPU 4Ghz Test A
    306.23 HPET-ON CP-Off CPU 4Ghz Test A

    305.53 HPET-Off CP-On CPU 4.2Ghz Test B
    306.23 HPET-ON CP-On CPU 4Ghz Test B
    306.23 HPET-ON CP-Off CPU 4Ghz Test B

    This one produced the most interesting/bizare results. (Which you can see from the varying numbers and most importantly the graphs and the shapes of the ridges)
     
  9. -Tj-

    -Tj- Ancient Guru

    Messages:
    17,209
    Likes Received:
    1,941
    GPU:
    Zotac GTX980Ti OC
  10. fellix

    fellix Member Guru

    Messages:
    194
    Likes Received:
    33
    GPU:
    KFA² GTX 1080 Ti
    Frame stuttering can't be evaluated in an FPS graph -- it's too coarse. Use the time-frame recording feature in FRAPS for this an plot the data in Excel. It would be interesting.
     

  11. Prophet

    Prophet Master Guru

    Messages:
    822
    Likes Received:
    14
    GPU:
    Msi 680 Gtx Twin Frozr
    http://www.evga.com/forumsarchive/fb.asp?m=100689230
     
  12. MrBonk

    MrBonk Ancient Guru

    Messages:
    3,265
    Likes Received:
    207
    GPU:
    MSI RTX 2080
    Oh well, I already ran the FFXIV benchmark, was gonna post it. But I ended up not doing it. The score stayed roughly similar though it took a small hit with Core Parking Off and my lowered CPU clock (Less than ~100 points though) Smoothness seemed about the same but the graph painted in the benchmark as it goes seemed like it had a lot more smaller fluctuations with HPET On, CP off and 200Mhz less CPU lock. But you couldn't really feel them.


    I also have RE5 Bench, I'll run it and see what happens. I won't go back and do 305.53 tests but I remember running it on that driver a little while back with CP on and HPET off.

    I get the impression of the stuttering from the sudden hard edge drop offs and spikes compared to the mostly smoothed and rounded edges off HPET on/CP-Off.

    I don't have Excel, but I do have fraps .. maybe i'll give it a try
     
  13. -Tj-

    -Tj- Ancient Guru

    Messages:
    17,209
    Likes Received:
    1,941
    GPU:
    Zotac GTX980Ti OC

    FF XIV is cpu dependent, and i noticed it can have some deviations as well, i never tested with HPET- off.
    I noticed fluctuations only when i changed drivers & changed in nvcpl FFXIV profile threaded optimization - on, triple buffer on, 2 frames to render ahead.

    If both enabled i got higher lowest fps (1-3fps).

    And yes stable cpu OC helps too with less fps drops. Just like by Re5 - fixed scene, its when i changes render camera angles (scene), by each change it can have fps drops or stay smooth.

    Anyway overall i would leave HPET - ON. :)
     
  14. Varna

    Varna Active Member

    Messages:
    92
    Likes Received:
    0
    GPU:
    EVGA 690GTX
    Thanks. LatencyMon does a good job of showing you which items are causing high DPC spikes too.

    [​IMG]

    It's not consistent at all. Jumps around from 40~600 just sitting idle.

    [​IMG]

    And here are the main culprits. I figure I can't really do anything about the nvidia drivers. But the other items? Any tips?
     
  15. -Tj-

    -Tj- Ancient Guru

    Messages:
    17,209
    Likes Received:
    1,941
    GPU:
    Zotac GTX980Ti OC

  16. Prophet

    Prophet Master Guru

    Messages:
    822
    Likes Received:
    14
    GPU:
    Msi 680 Gtx Twin Frozr
    I disagree.


    Btw, although Tj´s way is good, this seems a little bit more complete (and it has worked for me in the past).

    http://img571.imageshack.us/img571/6330/intelshortcut1.png
     
    Last edited: Sep 29, 2012
  17. Varna

    Varna Active Member

    Messages:
    92
    Likes Received:
    0
    GPU:
    EVGA 690GTX
    All my chipset drivers are up to date... no improvement.

    So where exactly do I get that xperf application? The main page always leads me to a dead end... don't see a download link.
     
  18. Varna

    Varna Active Member

    Messages:
    92
    Likes Received:
    0
    GPU:
    EVGA 690GTX
    I followed the guide you posted Prophet and I did get a lot of the same results... but a few new ones worth looking into. Now the problem is I don't really know where to go from here. The main culprits seem to be..

    USBPORT.SYS
    Wdf01000.sys
    ataport.SYS
    cmudaxp.sys
    ndis.sys
    ntoskrnl.exe
    tcpip.sys
    nvlddmkm.sys
    tunnel.sys

    So basically a ton of stuff. Is there any way to narrow this stuff down anymore?
     
  19. Cyberdyne

    Cyberdyne Ancient Guru

    Messages:
    3,582
    Likes Received:
    302
    GPU:
    2080 Ti FTW3 Ultra
    DPC can help narrow problems, true. But I think you guys are stressing over it way too much.
     
  20. Varna

    Varna Active Member

    Messages:
    92
    Likes Received:
    0
    GPU:
    EVGA 690GTX
    No doubt I'm pretty much obsessing over it at this point. I started looking into it wen someone over at the Guild Wars 2 forums said high DPC latency could cause performance issues in such a CPU intensive game. I have a 690GTX and an i5-2500k @4.8GHZ and it should have no problems running the game... but it is. Severe problems. Stuttering mainly. So I just want to get this sorted out.

    Considering all the supposed problem areas my system has. Would I be better off just doing a new system install and taking it from there?
     
    Last edited: Sep 30, 2012

Share This Page