Discussion in 'Videocards - NVIDIA GeForce Drivers Section' started by Bukkake, Sep 18, 2012.
Do you use external (USB) disk?
Nope. I've checked to ensure all my chipset drivers, lan drivers and mouse and keyboard are up-to-date as well. I've tried with Hpet on/off, completely disabling my CPU OC, etc.
Btw if you have powersaving features enabled your dpc will be high. Ifyou disable them or even run something as prime95 they will be lower.
Which ones? I don't really want to turn off the one in the bios that makes your CPU constantly run maxed out.
Well damn. Looks like there is nothing I can do about it.
I uninstalled my Nvidia drivers just to see and yup, all the latency was gone. 50~70 DPC.
Once I installed it again it went up again to 70~600 DPC. I have my GPU installed on a PCI-E 2.0 board... could that have anything to do with it?
Funny thread really, lol. These things are best left alone.
It's a PC enthusiast forum man... always something new to obsess about.
I'm not even getting audio pops, hiss or interruptions. I'm only getting some bad stuttering in Guild Wars 2 and something that appears to be micro-stuttering in Borderlands 2 (though the devs acknowledge something funny going on with 680s).
I'll give up on this. Seems uninstalling my 690GTX drops all the heavy spikes.. no way i'm replacing my gpu though.
Open your DPC latency monitor, then open your Nvidia Control Panel and go to 3D Settings->Adjust image settings with preview.
Does the latency value change when the 3d image is spinning?
The range seems the same. 60~600.
I've come to the conclusion that 306.23 ,Core Parking Off and HPET (or an unstable OC) just do NOT play well together with FFXIV. I get intense microstuttering on every combination possible aside from HPET OFF and stable 4Ghz with CP on or OFF (Even then there is some still minor Microstuttering near the end of the bench)
So i'm just chucking it up to the drivers. I never had these issues with 305.53.
I have done benches with varying configurations of RE5 , I will edit those in this post soon.
I agree after trying this.
This has been answered HPET Give different results for different people different hardware and different OS
Infact this was said in the first page and on the links. HPET is depended on OS/Hardware and everything in between including the makers of said parts
Or was different results for different OS and different Hardware not good enough of an answer
People general dont answer something that has already been answer and stated in the thread to being with.
Depends also on how HPET is implemented on the motherboard, if its 32 or 64 bit, if it depends on APCI and what kind of software your using.
Gaming FPS difference isn't much thou, merely 1-3 fps differences between HPET on or off, if that value is greater something else is wrong that most probably causes out of sync issues, but none the less, HPET solves even those issues, I just think it's the wrong approach to force a timer on a driver rather than getting or writing a driver that makes use of HPET.
There are also people reporting that they got rid of stuttering by turning HPET off in their BIOS, but that might as well be a poorly implementation of HPET on their motherboard.
From my experience, using BIOS default with HPET enabled and forcing Windows to use HPET only works best overall.
By default Windows 7 uses different CPU timers to calculate.
HPET is the newest and best of these timers, but because of the default combination of timers, it takes longer time for the CPU to sync between them.
Which also can result in clock offsets, especially if you are overclocking.
Even Linux has benefits from HPET btw.
HPET is also the timer with the lowest DPC latency by nature, since it is fast and doesn't need to sync, but that can probably cause issues with older software that expect the timer to sync.
There is a lot of reading about HPET tuning over at IBM, but gets a bit nerdy and is ofc. all about server hard and software.
i see no lose or gain with hpet other then 14mhz timing now, and slightly smoother mouse control.
And any games I had that do stutter have abouslty nothing to do with hpet on or off or any combination of it. So I put it on 64bit bios and on in win7 and left it.
Bottom line it all dependent on software hardware and who made the hardware, as far I can tell there is no standard to implementing it on hardware which probably why it is hardware and brand/manufacture dependent.
I would agree with this 100%. I have tried all sorts of tests and I agree.
So I'm still using the old faithful vista home premium 64-bit with sp2 because i don't see a damn performance difference between it and windows 7, but i'd like to ask, how does one disable HPET in Vista?
When i type bcdedit /set useplatformclock true in the command prompt (with admin rights) it tells me value i have entered is invalid and doesn't exist. This works in windows 7 and later is what it seems, but vista may be lacking the option.
I ran wintimertester 1.1 and my queryperformance frequency was 25.00000mhz which seems like its related to the Fsb setting like the first page of this thread mentioned. So if its at least 10mhz or higher i heard it was enabled and is this so in my case?
I have the 790i SLI FTW PWM from Evga as specified to the left and the BIOS has a straight hpet function=enabled/disabled option so i know the BIOS supports it.
Vista should be fine
There are 4 possible combinations of this by changing HPET on your mobo and in windows.
I have found this on the net some time ago, and saved it for later reference.
"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."
You might wanna try all 4 options, see wich one works best.
It seems option **** works best, as it has HPET enabled in the bios and windows.
Did some follow up testing to see if the difference in latencymon readouts translates to any palpable change on my rig. results are completely identical.
By HPET on i mean both windows and BIOS settings. FPS readouts from 3 loop fraps runs following the same path as humanly as posible.
OS boot time 19s
avg. game load 7s
OS boot 20s
avg game load 7s
I noticed this also, but i though i was imagining it.