My Gigabyte Gtx 1080 Xtreme - till today stuck on 1278mhz clock. Help.

W

wes28222

Limp Gawd
Joined
Jul 20, 2016
Messages
134
Help i dont know what is causing this. So till today was everything working fine. Boost in 3d was 2000mhz in core clock. Today i run some game i saw that clock is very very below normal values.
So like i said.
I didn't install any software IE: MSI afterburner, precision or anything like that. I dont change anything in Windows and bios.
Even dont change power managament in windows. All default like always.
This issue its from itself. I am using 368.692. 2 weeks card boosting fine to 1990-2000mhz. But just run today any game and i saw low fps, check gpu core clock. And now i have 1278mhz no matter what.

Look:
1080b.jpg




PC:
6700K no OC
2x8GB Kingston Hyperx
Corsair 750 RM
Gigabyte Gtx 1080 Xtreme
SSD Crucial Bx 100 256 GB
Asus Z170-P
Windows 10


Is my hardware GPU failing or psu or something other?
 
If you download and open MSI afterburner, what's the power limit set at? Can you push it up to 100%?

Also, check the power cords going to the card?
 
Yes power cords are going to the card,because there are no signal from leds that is something wrong.
 
Your screen shot shows it being capped by the power limit.
 
Pwr limit had always even with 2000mhz boost. Its not this. 2 weeks was fine. Until today i check gpuz and core clock is 1278mhz . I dont chang anything in Windows and bios.


Something change from itself or some part of pc is failing.
 
as you are using FRAPS, check if by error didn't you enabled the Frame Limiter option..
 
Hmmm i made restart pc and all is fine clocks normally 1990-2000mhz. ANy ideas why that happened? Hardware possibly or no?
 
wes28222 said:
Hmmm i made restart pc and all is fine clocks normally 1990-2000mhz. ANy ideas why that happened? Hardware possibly or no?
Click to expand...

no, absolutely not. software problem or bug.. probably FRAP Frame Limiter option was turned ON by error, it's common..
 
After restart its fine ,clocks 1990-2000mhz instead of 1278mhz and working like it should and fps is normal. So propably bug no hardware issue? No to worry?
 
Often downclocking is associated with overheating, bad drivers, unstable power and sometimes faulty hardware. That somebody said to me. Maybe with hardware something that was?:)
 
Last edited:
Often downclocking is associated with overheating, bad drivers, unstable power and sometimes faulty hardware. That somebody said to me. Maybe with hardware something that was?
 
I had the same issue happen when I turned on adaptive vsync in nvidia control panel. I forgot I did that and then I ran 3d mark I got artifacts, corruption and my card was capped at like 1000mhz core speed. I was like WTF, then I realized I had turned that on adaptive vsync in control panel in the global settings page. Turned it off and all was good.
 
No i dont change VSYNC options . Its not this.

There are few peoples on GF forum with the same issue. I dont know if it is related to hardware or not. Card on load with normal boost 1990-2000mhz is stable , no crashes, no freezes.

And it doesnt happen under load. Just when the PC has been idle for some hours. Then the boost clock speed gets stuck at 1278 Mhz until I reboot my PC and then it's at full boost clock speed again.
GPU-Z reports the "perfcap" reason as PWR.

It happened just once and i dont know why.
 
I've been having the same issue with my 1080FE since day one. And, yes, quite a few people are reporting similar issues in the Geforce forums. There are also reports of GTX9xx users experiencing the issue and have been since last summer.

Card runs absolutely fine with no stability issues, even under heavy load for extended periods. But at some point and time when the card is idling, doing nothing, the clock speed will gimp and cap out at much lower than it should. Mine has capped as low as in the 800s and as high as the 1500s. Memory clock speed is not affected. There are no reported driver crashes, nothing in the logs. Applications and games will still run, just slower than they should. The problem never occurs while "in use". For me, the problem occurs as frequently as once or twice a day to as little as once a week. I leave my machine on 24/7, no sleep, no hibernation, power options in the OS set to High Performance. The monitor is the only thing set to go to sleep after a set time (turning this setting on or off doesn't make a difference, tried it). I've reviewed the P-state with Nvidia Inspector and that isn't hanging up. I've tried with G-sync on and off with no difference.

At first I thought a reboot was needed to "reset" the issue. But if you are using Afterburner, you can create profiles. When the problem acts up, just reapply the profile and it will working great again as if nothing happened.

I've updated drivers, done clean uninstalls with DDU, etc., no change. The computer is all new, built in April. I didn't experience these issues with my GTX970, but it was also using older drivers.
 
I was
VirtualMirage said:
I've been having the same issue with my 1080FE since day one. And, yes, quite a few people are reporting similar issues in the Geforce forums. There are also reports of GTX9xx users experiencing the issue and have been since last summer.

Card runs absolutely fine with no stability issues, even under heavy load for extended periods. But at some point and time when the card is idling, doing nothing, the clock speed will gimp and cap out at much lower than it should. Mine has capped as low as in the 800s and as high as the 1500s. Memory clock speed is not affected. There are no reported driver crashes, nothing in the logs. Applications and games will still run, just slower than they should. The problem never occurs while "in use". For me, the problem occurs as frequently as once or twice a day to as little as once a week. I leave my machine on 24/7, no sleep, no hibernation, power options in the OS set to High Performance. The monitor is the only thing set to go to sleep after a set time (turning this setting on or off doesn't make a difference, tried it). I've reviewed the P-state with Nvidia Inspector and that isn't hanging up. I've tried with G-sync on and off with no difference.

At first I thought a reboot was needed to "reset" the issue. But if you are using Afterburner, you can create profiles. When the problem acts up, just reapply the profile and it will working great again as if nothing happened.

I've updated drivers, done clean uninstalls with DDU, etc., no change. The computer is all new, built in April. I didn't experience these issues with my GTX970, but it was also using older drivers.
Click to expand...
I was thinking that this is maybe because i have monitor setting to turn off after 15 minutes. But like you said its not this issue. Maybe something with pcie power management or something? Or maybe faulty card?
 
If you have your power options set to Full Performance, then PCIE power management shouldn't be powering down the card. If the card was faulty, as I mentioned in the other forum, you would expect it to have issues running under full load where temps, power, and tolerances can be a lot more temperamental.

As experienced, the card nor its drivers never crash. Applications still work fine and don't bomb out either. It just runs slower until you "refresh" its settings, whether it be from a reboot or reapplying a profile.

I'm leaning more towards a driver issue (GPU or another piece of hardware interfering), a Windows configuration problem (be it from an update, file change, setting, etc.), or a combination of the two. I guess there could be a slim chance a compatibility issue with a motherboard or a BIOS configuration could be the problem, but I haven't seen anything in my setup that would leave me to believe it is that (at least with my situation).

Since it isn't crashing, there is nothing in the logs to point us in the right direction. If we were able to somehow enable some kind of advanced logging within the driver that could track changes, then that could be helpful. But I don't know how to do that.
 
I left pc to another 3 days in idle, this time clocks was normal in 3d, no throttle. So it happened once just. Or maybe i must left idle longer time. Dont know.


Also many peoples have this issue on geforce com so i think its not hardware related.
 
I have something:

GTX 980 Ti occationally gets stuck at idle clock speed


And reason was oc card.



"When my driver crashes, it usually tells you with it pop-up saying that the driver has crashed right? because mine doesn't do that anymore, but with my old 970, it would black screen and then tell me the driver has stopped and recovered but like i said earlier, now it just gets stuck at 405 when i'm on my desktop browsing chrome or something. " ( even without driver crasH )
 
wes28222 said:
I have something:

GTX 980 Ti occationally gets stuck at idle clock speed


And reason was oc card.



"When my driver crashes, it usually tells you with it pop-up saying that the driver has crashed right? because mine doesn't do that anymore, but with my old 970, it would black screen and then tell me the driver has stopped and recovered but like i said earlier, now it just gets stuck at 405 when i'm on my desktop browsing chrome or something. " ( even without driver crasH )
Click to expand...
It doesn't explain why it happens to people running stock.

I think you might be trying too hard trying to find every slight issue and piecing it together as if it is related to this. You keep bouncing back and forth from one theory to the next even after there have been explanations stating why one might be possibility or why one might not be.
 
Oh is this the bug where the GPU gets stuck at a specific clock after a long time idling with the display turned off? If so there is an easy fix, download nvinspector, open overclocking tab, and click "make a shortcut", then just use the shortcut everytime the bug occurs.You can also click 'apply defaults' of course.
This makes the GPU runs at its default settings again (obviously if you are doing software OC'ing this will reset it...).

This is the easiest way I can fix it on my 980 ti (not OC'd btw, just factory OC, I even tried with a reference 980 ti BIOS and the bug still happens - it's a pure driver thing). Other solution is of course rebooting the PC but no thanks. Oddly enough even disabling/enabling video adapter doesn't cut it.
Not sure when this started happening, if you revert to old drivers you can fix it but you'll break some recent games like Doom. And actually a Pascal card might not like those old drivers :)

I thought this bug only happened with Maxwell cards though... Guess I'll have to live with it, it's been reported a long time ago already. My PC is on 365/24/7 but I like to turn off the display completely rather than let it go to sleep (in the Windows settings I mean), not sure if that matters.

Funny thing is sometimes my card gets stuck on its highest boost clock and won't throttle as it should when getting hot or going over TDP. Pretty scary but I just use the nvinspector trick every day when I get back from work before I even attempt gaming. I'm also always monitoring clocks/temp/TDP anyway.
Usually it just gets stuck in a low boost clock state though, noticeably hurting performance (that's how I found out the first time).

Also I'm on Windows 7 so not an OS bug, it's purely on nvidia's side.

edit : see this long-ass thread that went from 970 to 1080 : https://forums.geforce.com/default/...0-clock-rate-throttling-and-temporary-fix-/1/
Apparently nvidia is aware but can't reproduce it. Urgh.
 
Last edited:
kalston said:
Oh is this the bug where the GPU gets stuck at a specific clock after a long time idling with the display turned off? If so there is an easy fix, download nvinspector, open overclocking tab, and click "make a shortcut", then just use the shortcut everytime the bug occurs.You can also click 'apply defaults' of course.
This makes the GPU runs at its default settings again (obviously if you are doing software OC'ing this will reset it...).

This is the easiest way I can fix it on my 980 ti (not OC'd btw, just factory OC, I even tried with a reference 980 ti BIOS and the bug still happens - it's a pure driver thing). Other solution is of course rebooting the PC but no thanks. Oddly enough even disabling/enabling video adapter doesn't cut it.
Not sure when this started happening, if you revert to old drivers you can fix it but you'll break some recent games like Doom. And actually a Pascal card might not like those old drivers :)

I thought this bug only happened with Maxwell cards though... Guess I'll have to live with it, it's been reported a long time ago already. My PC is on 365/24/7 but I like to turn off the display completely rather than let it go to sleep (in the Windows settings I mean), not sure if that matters.

Funny thing is sometimes my card gets stuck on its highest boost clock and won't throttle as it should when getting hot or going over TDP. Pretty scary but I just use the nvinspector trick every day when I get back from work before I even attempt gaming. I'm also always monitoring clocks/temp/TDP anyway.
Usually it just gets stuck in a low boost clock state though, noticeably hurting performance (that's how I found out the first time).

Also I'm on Windows 7 so not an OS bug, it's purely on nvidia's side.

edit : see this long-ass thread that went from 970 to 1080 : https://forums.geforce.com/default/...0-clock-rate-throttling-and-temporary-fix-/1/
Apparently nvidia is aware but can't reproduce it. Urgh.
Click to expand...
If you have Afterburner installed, you can also just create a profile and apply the profile when the problem occurs.

I never leave my monitor on all the time, so I can't verify if it is when the monitor powers off. But I have tested it with both the monitor set to go to sleep and no monitor power off, with the monitor to set to go to see and monitor powered off, and with the monitor not set to go to sleep and the monitor powered off. All three results experience the same issue.

I will say that initially I thought it was the monitor sleep mode causing the problem because when I turned that setting off I didn't have an issue for just a little over a week. But when the issue came back even set that way and began happening again at the more frequent intervals, I figured that wasn't it.

Another thing I will say is that when I do turn my monitor on, the desktop will appear on the display but then sometimes I will get a single flicker within 1-3 seconds of it turning on as if the screen is achieving connectivity again. Also, every time I turn off my monitor I hear the hardware disconnect tone Windows loves to play when you unplug devices from your machine and will hear the hardware connected tone when the monitor gets turned on. I don't know if it does it when the display goes to sleep since I am never around when that happens. My older display never did that, but it was connected via HDMI while my new monitor is connected via DisplayPort. It's a brief, but minor, annoyance and I've always made a mental note wondering if that might be playing some part, or at least providing a hint, as to what the culprit to this problem might be.
 
I left pc on idle 4-5 days and problem with throttled clock due to PWR limit started. Restart pc to fix this. I left again pc but this time for 3 days in idle and nothing happened. So i think 3 days are not enough
smile.gif
 
You must log in or register to reply here.
Back
Top