TWEAK: Enable HPET (in BIOS and OS) for better performance and FPS


Recommended Posts

By default Windows 7 uses different timers in the CPU to calculate stuff. HPET is the newest and best of these timers, but because of default combination of timers it takes longer time for CPU to keep up all the timers and sync between them. Forcing Windows to use HPET only improves performance and leads to greater FPS.

Steps to enable this tweak:

1. Enable HPET in BIOS. If you have HPET option in BIOS then your hardware can support HPET.

2. Enable HPET in Windows by giving this command in admin credential CMD:

bcdedit /set useplatformclock true

3. Reboot

  • Like 2

Enabling HPET in BIOS is just half way of enabling HPET, it needs to be enabled in OS too, and in a way that it's the only timer used.

By default windows uses combination of TSC+ACPI timers, not matter if HPET is enabled in BIOS.

TSC+LAPICs Low performance (slow timers + syncing)

LAPICs low performance (slow timer - no syncing)

TSC+HPET medium performance (slow and fast timer + syncing)

HPET high performance (fast timer - no syncing)

HPET + platformclock=true will give you best timer resolution, frame rate and lowest DPC latency.

You can test timer ratio and QueryPerformanceFrequency with WinTimerTester 1.1 http://www.mediafire...xzo9n84d8lze9nb

The higher the QueryPerformanceFrequency is the better is performance. You only get high frequency with HPET. The other timers will give you significantly less frequency. Also note that if your ratio is not 1.0000 you are off set (or you have wrongly OC'ed), enable HPET and you should be without sync problems.

If you ever want to go back to default timers admin cmd:

bcdedit /deletevalue useplatformclock

Varying depending on setup, one should get increase up to +30 FPS and from the between.Online games is a good example of boost from HPET.

TSC+LAPICs Low performance (slow timers + syncing) = 2.76MHz

LAPICs low performance (slow timer - no syncing) = 3.5Mhz

TSC+HPET medium performance (slow and fast timer + syncing) = 3.8Mhz

HPET high performance (fast timer - no syncing) = 14.3MHz

Run the WinTimerTester 1.1 to see your QueryPerformanceFrequency

Then try with HPET, you'll be amazed.

Enabling HPET in BIOS is just half way of enabling HPET, it needs to be enabled in OS too, and in a way that it's the only timer used.

By default windows uses combination of TSC+ACPI timers, not matter if HPET is enabled in BIOS.

TSC+LAPICs Low performance (slow timers + syncing)

LAPICs low performance (slow timer - no syncing)

TSC+HPET medium performance (slow and fast timer + syncing)

HPET high performance (fast timer - no syncing)

HPET + platformclock=true will give you best timer resolution, frame rate and lowest DPC latency.

You can test timer ratio and QueryPerformanceFrequency with WinTimerTester 1.1 http://www.mediafire...xzo9n84d8lze9nb

The higher the QueryPerformanceFrequency is the better is performance. You only get high frequency with HPET. The other timers will give you significantly less frequency. Also note that if your ratio is not 1.0000 you are off set (or you have wrongly OC'ed), enable HPET and you should be without sync problems.

If you ever want to go back to default timers admin cmd:

bcdedit /deletevalue useplatformclock

Varying depending on setup, one should get increase up to +30 FPS and from the between.Online games is a good example of boost from HPET.

A guy on another forum claimed that by disabling it, his DPC latency plummeted and his gaming performance improved; most notably the microstuttering.

So I tried it myself, and the difference to my astonishment, was very noticeable. First off, my DPC latency dropped big time. Not that it was high to begin with, but now it hovers below 10 for the most part.

Second, I noticed that general windows performance felt snappier. I know I'm not imagining things, because like many, I scrutinize my system so I'm sensitive to even minute changes in performance.

And last but not least, gaming performance is definitely smoother. Anyone with SLi will tell you that occassionally when playing a game, you experience a bit of lag or stuttering for no apparent reason whatsoever. Most people attribute these minor discrepancies to SLi, but what if it's something else? Because I can tell you, that my overall gaming experience has improved since turning HPET off.

If you have the HPET option in your BIOS, I highly recommend experimenting with it and see if it impacts your performance. There's a possibility that having it on could increase the amount of microstuttering in any given game; particularly if you're running SLi.

http://forums.nvidia...howtopic=183329

Without changing any settings in the BIOS or the OS, and with HPET enabled in the BIOS as normal I get this

Untitled.png

^Because if you enable HPET in BIOS you will mix it with other timers and CPU needs to sync between them. You need to force HPET only and reboot. Mixing HPET with other timers is not good, but using HPET only will give you best performance. People don't know about that Windows needs to be forced to use HPET only - that's why the bios setting alone can undergrade their performance.

bcdedit /set useplatformclock true (then reboot) enable HPET

bcdedit /deletevalue useplatformclock (then reboot) disable HPET

Without changing any settings in the BIOS or the OS, and with HPET enabled in the BIOS as normal I get this

Untitled.png

You are using TSC+HPET. Try with HPET only (bios and OS setting) you should get something like 14.xMhz for queryperformancefrequency.

^Because if you enable HPET in BIOS you will mix it with other timers and CPU needs to sync between them. You need to force HPET only and reboot. Mixing HPET with other timers is not good, but using HPET only will give you best performance. People don't know about that Windows needs to be forced to use HPET only - that's why the bios setting alone can undergrade their performance.

bcdedit /set useplatformclock true (then reboot) enable HPET

bcdedit /deletevalue useplatformclock (then reboot) disable HPET

Ok so I enabled it, rebooted

2.PNG

First things I notice are it took twice as long to finish loading the desktop and gadgets

Mouse pointer has a ghost trail now like I am using a cheap LCD with 20ms response rate

What exactly is "queryperformancefrequency" ? and how does it effect the OS ?

queryperformancefrequency is frequency of high resolution timer performance counter, timers are different as I have explained above, and better/higher the timer frequency the better responsive the system is. With HPET enabled in BIOS and OS you can achieve highest timer performance resluting in better responsive system.

This "tweak" is depending on each hardware setup. Some people say it slightly speeds things up, others start having micro-stutter in multimedia/3d gaming, especially SLI setups. the DPC latency goes through the roof and limits the FPS in the games. Best thing to do is try it for yourself, do some benchmarks and see what happens, if your system runs fine now. Leave it alone would be my advice.

Enabling HPET in BIOS mixes HPET with other timers - needing the CPU sync between the timers resulting in compromised performance. If WIndows is forced to use HPET only (in bios and os) it in modern system results in higher performance, higher FPS and better DPC.

Test and try what suits you.

Enabling HPET in BIOS mixes HPET with other timers - needing the CPU sync between the timers resulting in compromised performance. If WIndows is forced to use HPET only (in bios and os) it in modern system results in higher performance, higher FPS and better DPC.

Test and try what suits you.

I had mouse ghosting and other nice things when doing it "your" way. Disabling HPET is the best way for me on a Sandy Bridge-E CPU and an ASUS Sabertooth motherboard.

commands:

bcdedit /set useplatformclock true (then reboot) enable HPET

bcdedit /deletevalue useplatformclock (then reboot) disable HPET

TSC+LAPICs (useplatformclock false)*

LAPICs (useplatformclock true)**

TSC+HPET (useplatformclock false)***

HPET (useplatformclock true)****

* Windows default.

**platformclock=true and HPET disabled in BIOS will default to LAPICs, which is good compared to TSC, but doesn't not have not so high resolution and so low DPC latency as HPET.

*** Windows default with HPET enabled in BIOS.

**** HPET enabled in BIOS and in OS.

It seems to have had the opposite effect here too. I had HPET enabled in the BIOS already but hadn't set the useplatformclock boot parameter.

Without it, my clock ratio was mostly 1.0000, occasionally flickering to 1.0001. After setting it my clock ratio is mostly 0.9999, occasionally 1.0000.*

Here's the original post from the author of WinTimerTester 1.1. It explains the scenario better. Seems to be related to software overclocking utilities.

The problem is that for many computers, this function in Windows 7 is now based on the clock speed of the processor. If you overclock in the bios and then boot up, this timer is calibrated and will work 100% correctly. If you use SetFSB or a similar program and you change the bus speed from within Windows, you have now just screwed up this very important timing mechanism. The number of applications that can choke after you do this is surprising, especially games.

* After a few minutes it seems to have settled at 1.0000. I don't have any mouse ghosting.

I had mouse ghosting and other nice things when doing it "your" way. Disabling HPET is the best way for me on a Sandy Bridge-E CPU and an ASUS Sabertooth motherboard.

Funny as I don't have any problems like that. my CPU is i5 3570K in Z77 mobo and Asus HD 7850. I do use CRT monitor.

Perhaps it's something else you need to adjust if you want to have best timer.

Well RAGE was smooth when it ran, but every few seconds was locking up for a few seconds

BF3 let me play it with CFX for the first time so it seems to have helped there

I`ll test a few more games but it seems to have negatives and positives for different things so far

I've read many people with the Intel X58 chipset have issues with HPET, checked my BIOS and it was enabled (in 32-bit mode yet I run a 64-bit system). Disabled it because lately my computer has hard locked twice while browsing the internet (hasn't done it in awhile though). So maybe my X58 chipset + HPET in 32-bit was part of the issue. Either way my system seemed more responsive on boot than usual, won't be able to tell if it helped until I get a chance to use it more or see if it freezes for no reason again.

I've read many people with the Intel X58 chipset have issues with HPET, checked my BIOS and it was enabled (in 32-bit mode yet I run a 64-bit system). Disabled it because lately my computer has hard locked twice while browsing the internet (hasn't done it in awhile though). So maybe my X58 chipset + HPET in 32-bit was part of the issue. Either way my system seemed more responsive on boot than usual, won't be able to tell if it helped until I get a chance to use it more or see if it freezes for no reason again.

I have X58 chipset boards and never saw an issue with 64bit HPET

Funny as I don't have any problems like that. my CPU is i5 3570K in Z77 mobo and Asus HD 7850. I do use CRT monitor.

I have 5?s latency with HPET disabled and 60?s with it enabled using your method, plus no mouse ghosting or stuttering which I get with it enabled. There is nothing else concerning this that I can 'adjust'. I guess my X79 board just feels better with it disabled or something.

Is this on TSC+LAPICs?

wintimertester.png

Plus I'm unsure if the Asus P7P55D Pro has a HPET toggle, will have to check later.

Hard to tell, it's either TSC+LAPICs or LAPICs. Your ratio is excellent so it might indicate that there is no sync - meaning you might have LAPICs. However hardware differs and even if your ratio is that exact in 1.0000 it still might sync TSC+LAPICs well - depending on hardware. Windows obviously uses LAPICs mainly for hypervisor/virtual machine and TSC as general, but it's a bit mystery what Windows 7 uses, a combination is the best assumption. Vista was heavily pushed into using HPET. I would say that you have at least TSC. If you want TSC + HPET enable HPET in bios. And if you want to test HPET only use the bcdedit /set useplatformclock true.

I think that your mobo has HPET in BIOS as HPET isn't actually anything new.

-----------------------------

Good to hear that people are getting good results with HPET (bios & os). However I have provided details about timer settings in the first page and how to revert back from HPET.

Hard to tell, it's either TSC+LAPICs or LAPICs. Your ratio is excellent so it might indicate that there is no sync - meaning you might have LAPICs. However hardware differs and even if your ratio is that exact in 1.0000 it still might sync TSC+LAPICs well - depending on hardware. I think that your mobo has HPET in BIOS as HPET isn't actually anything new.

Couldn't find anything in the BIOS, think it's hidden. Though I ran bcdedit /set useplatformclock true and I'm getting 14.31818 MHz now. But the ratio took longer to stabilise at 1.000 (took 180 seconds), it was alternating between 0.9997 and 0.9999 a lot.

Not spotted anything bad yet, like ghosting mouse pointer or such. My DPC latency remains unchanged, averages at 60-80us

Im running it on a i7 [email protected] (clocked in BIOS), and Asus P8Z68-V Pro mobo/NV 570GTX and enabling HPET & setting windows to solely use HPET via command prompt has fixed my micro-stuttering id see in BF3.

Result :) Cheers OP I owe ya a beer :)

This topic is now closed to further replies.
  • Recently Browsing   0 members

    • No registered users viewing this page.
  • Posts

    • I've been using MWB Premium for a number of years so that along with Windows updates and updated browser should be fine. Thanks for that.
    • For some reason I suddenly have the urge to go shopping at Sears.
    • So I did a quick test based on 3+ different public instances from the litany at searx.space ... and it spins everything rather differently. It seems that SearXNG is a meta-search engine (queries multiple search indexes rather than only Google's or Bing's or Wikipedia's or Reddit's) that operates in two modes: > public instances ... each instance opens itself to outside users who piggyback on its cached search history; this instance's own identity becomes known/tracked but end-users are hidden similar to an anonymization proxy; this instance's querying of major search indexes may be API based [rated limited, blocked, etc.]). > private instances ... your private install/instance that itself queries multiple (configurable) search indexes of crawled web content; every major Search Engine associates all traffic to your private instance (so your traffic is tracked via network usages) but client-side tracking (your own browser/computer specs) is flushed because it's a "server" doing the querying rather than your browser. My test asked the same 1 question to the 3+ engines and they all returned vastly different results: some had CAPTCHA failures against Google, some had failures against Wikipedia, and the actual results were also different -- some had auto-complete enabled, others returned a wikipedia highlighted excerpt despite the Wikipedia failure (hinting at results being cached from previous keyword matching), and others just gave an Are-You-Human non-CAPTCHA loop before returning random results. So this begs the caveat: Search query results will vary based on which instance is used because every instance queries the other search indexes separate (and thus its results are influenced on that instance's aggregate search history and index-access limitations). The major distinctions for SearXNG versus DDG or Brave: > The search UI is 'untracked' since no UI trackers are baked-in which would phone home or lay cookies into your browser (for DDG/Brave usage stats), > There is no 'crawler' that canvasses the Internet to discover fresh content (it leaves that to the major search indexes), > Queries multiple search indexes ("meta-search engine") based on the configurations and usage history of the server instance, > Privacy-friendly due to its ability to shield user tracking via standing up a non-local server instance connectable to major VPN providers: queries would all appear to come from general VPN/Proxy providers rather than your private instance (whether installed locally or on your own VPS in the cloud). PS: I've previously come across specialized search engines of this nature that indexes searches across media assets like YT, OF, etc. SearXNG seems to be a good backbone...if the rate-limiting/captcha/etc. issues were resolved.
  • Recent Achievements

    • Week One Done
      flexorcist earned a badge
      Week One Done
    • One Month Later
      Woland13 earned a badge
      One Month Later
    • Week One Done
      Woland13 earned a badge
      Week One Done
    • One Year In
      bernmeister earned a badge
      One Year In
    • Week One Done
      Scoobystu earned a badge
      Week One Done
  • Popular Contributors

    1. 1
      +primortal
      490
    2. 2
      +Edouard
      225
    3. 3
      PsYcHoKiLLa
      147
    4. 4
      Steven P.
      74
    5. 5
      FloatingFatMan
      71
  • Tell a friend

    Love Neowin? Tell a friend!