Overclock.net banner

GTX 1080 TI GPU Frequency changing by itself

4K views 12 replies 6 participants last post by  mamuf 
#1 ·
Hello.

I just bought a ZOTAC 1080 TI AMP EXTREME and I'm having some issues with it.

I had some BSOD after installing the graphic card. So I uninstalled all Nvidia drivers and softwares and did a clean install. Until now, all good. So I decided to test it using FurMark v1.19 and Realbench v2.54. While testing I noticed that the GPU frequency went as high as 2000MHz by itself on the Realbench test, from the start! The ZOTAC GPU software (FireStorm) is with its default config, so it shouldn't go pass 1759MHz.

I tried using MSI afterburn but I got the same issue.

Does anyone knows if this is normal? What can I do to prevent this?

Thank you
smile.gif
 
See less See more
1
#2 ·
going to steal this quote from reddit:
Quote:
First, the TLDR version:
GPU Boost 3.0 = GPU self overclocking. Card boosts to way beyond advertised clock due to available margins. It's normal, enjoy Pascal
smile.gif
nvidia has been doing this since kepler (the 600-700 series) with refinements along the way with version 3 now (kepler,maxwell and pascal). maybe do a little googling for "nvidia gpu boost" ?

edit:

you could edit the bios for both kepler and maxwell and change/control boost clock speeds. but pascal is locked down (w/some digital signature) so at this time no way to remove it.
 
  • Rep+
Reactions: CryWolf
#4 ·
Quote:
Originally Posted by Coopiklaani View Post

you should really be worried if the frequency doesn't change or doesn't go pass 1759MHz.
rolleyes.gif
The problem is that it passes, it goes to 2000MHz.
Quote:
Originally Posted by looniam View Post

going to steal this quote from reddit:
nvidia has been doing this since kepler (the 600-700 series) with refinements along the way with version 3 now (kepler,maxwell and pascal). maybe do a little googling for "nvidia gpu boost" ?

edit:

you could edit the bios for both kepler and maxwell and change/control boost clock speeds. but pascal is locked down (w/some digital signature) so at this time no way to remove it.
Nice! I didn't know that. I did some research on the gpu boost and looks like that this boosts up the gpu clock as high as possible within the temperature that you have and you set to be the maximum.

I only want to know, why the hell am I getting BSODs and my games are crashing. OverWatch crashed 4 times today so I just decided not to play it anymore since I solve this problem.

So the high gpu frequency is not the problem, right? What could it be?

Just a crazy question: My system is overclocked, I have a 6950x @ 4.0GHz / 4.1GHz with 1.20V and 4 memory sticks running at 3200MHz ddr4 (they were designed to run at this frequency). Before I had a GTX 1080 that was running pretty normal, no errors or BSODs after I stabilzed the overclock. Could the GPU change make the overclock unstable? If so, what would I have to change to stabilize it again?
 
#5 ·
Quote:
Originally Posted by CryWolf View Post

The problem is that it passes, it goes to 2000MHz.
Nice! I didn't know that. I did some research on the gpu boost and looks like that this boosts up the gpu clock as high as possible within the temperature that you have and you set to be the maximum.

I only want to know, why the hell am I getting BSODs and my games are crashing. OverWatch crashed 4 times today so I just decided not to play it anymore since I solve this problem.

So the high gpu frequency is not the problem, right? What could it be?
Welcome to the wonderful world of Nvidia drivers. Use DDU and download all the drivers compatible with the 1080ti. Install and test them until you find one that works. I feel your pain, I got mine to stop crashing but now I can't seem to get better than 85ms frametimes in all my games (which equals about 15fps)... my driver search continues.
 
#6 ·
Quote:
Originally Posted by bigjdubb View Post

Welcome to the wonderful world of Nvidia drivers. Use DDU and download all the drivers compatible with the 1080ti. Install and test them until you find one that works. I feel your pain, I got mine to stop crashing but now I can't seem to get better than 85ms frametimes in all my games (which equals about 15fps)... my driver search continues.
I'm well known of this world, lol. I can't count the times I had BSODs after updating drivers. But I can't believe that they didn't come up with a good driver for the GTX 1080 ti.... I can't even play my games without crashing ....
frown.gif


I had this lag problem that looked like 15fps while it was 150fps. I had to quit the game and run it again so it worked fine again.

I will try some drivers, if you find a good one, PLEASE, let me know!
biggrin.gif


Thank you.

Just one more thing: My system is overclocked, I have a 6950x @ 4.0GHz / 4.1GHz with 1.20V and 4 memory sticks running at 3200MHz ddr4 (they were designed to run at this frequency). Before I had a GTX 1080 that was running pretty normal, no errors or BSODs after I stabilzed the overclock. Could the GPU change make the overclock unstable? If so, what would I have to change to stabilize it again?
 
#8 ·
Quote:
Originally Posted by bigjdubb View Post

I would suggest returning everything back to stock until you figure out the crashing problem. It may not be the source of the problem but it helps to remove as many variables as possible.
Yeah, you are right. It has been 1 or 2 weeks that I'm using this overclock with no problem, and dam, it was hard to find the right tune. I will run some more test tomorrow with it at stock.

But hey, would the new graphic card mess up with my OC configuration? Would it demand more voltage from something that I could possible limited?
 
#9 ·
Quote:
Originally Posted by CryWolf View Post

I'm well known of this world, lol. I can't count the times I had BSODs after updating drivers. But I can't believe that they didn't come up with a good driver for the GTX 1080 ti.... I can't even play my games without crashing ....
frown.gif


I had this lag problem that looked like 15fps while it was 150fps. I had to quit the game and run it again so it worked fine again.

I will try some drivers, if you find a good one, PLEASE, let me know!
biggrin.gif


Thank you.

Just one more thing: My system is overclocked, I have a 6950x @ 4.0GHz / 4.1GHz with 1.20V and 4 memory sticks running at 3200MHz ddr4 (they were designed to run at this frequency). Before I had a GTX 1080 that was running pretty normal, no errors or BSODs after I stabilzed the overclock. Could the GPU change make the overclock unstable? If so, what would I have to change to stabilize it again?
Go back one or two driver updates and remove using DDU and follow the install I gave you in the other thread. Remember only install the graphic and physics driver using custom and clean install. The other Nvidia drivers will cause problems in games. Report back.
thumb.gif
 
#10 ·
Quote:
Originally Posted by bigjdubb View Post

I would suggest returning everything back to stock until you figure out the crashing problem. It may not be the source of the problem but it helps to remove as many variables as possible.
Quote:
Originally Posted by CptSpig View Post

Go back one or two driver updates and remove using DDU and follow the install I gave you in the other thread. Remember only install the graphic and physics driver using custom and clean install. The other Nvidia drivers will cause problems in games. Report back.
thumb.gif
So, I returned everything back to stock as bigjdubb suggested and tried to play some games, everything worked like a charm, so the new graphic card makes my overclock unstable. So I'm trying my same overclock with a little more voltage to see if works.

Right now I'm with 4.0GHz and 4.1GHz (core 6 and 10) @ 1.215V and cache at 3400MHz @ 1.15V, memory at 3200MHz @ 1.35V. My temperatures are the following:
Prime95 v28.10: Small FFTs: 87C - 20 minutes test
RealBench v2.43: It is crashing and giving me a BSOD at the beggining or after 30 minutes of test. I still don't know why. I think it doesn't support GTX 1080 TI. Sometimes Luxmark stops working too. Any ideas here?
X264 v2.06 (2 at the same time): 67C - 16 minutes
AIDA64 v5.7.38: 75C - 16 minutes test
OCCT v4.4.3 (Large Data Set): 80C - 15 minutes test
OCCT v4.4.3 (Small Data Set): 83C - 15 minutes test

**** UPDATE: I tried running RealBench after closing MSI Afterburner and other softwares from Razer and Corsair and it worked
biggrin.gif

RealBench v2.43: 68C - 15 minutes test

What do you guys think about the temperatures?
 
#11 ·
Quote:
Originally Posted by CryWolf View Post

**** UPDATE: I tried running RealBench after closing MSI Afterburner and other softwares from Razer and Corsair and it worked
biggrin.gif

RealBench v2.43: 68C - 15 minutes test
15 minute test is definitely not enough for stable OC. It's OK to find the limits. Try running RB for at least 2 hours. I had LuxMark crashing after 20-30 minutes (of course this does not say the CPU is unstable) and RB could BSOD even after few hours.

Also when you get BSOD, what does it say? If you miss the actual BSOD, go into Event Log under Windows Logs -> System and look for errors or even warnings. In my case, unless it's a WHEA error, LuxMark crashed after gpu driver failure which was logged only as a warning.

I am still trying to find out what causes the LuxMark crash. Already tried to close Afterburner (I use it to control gpu fan) and other programs like ASUS AI Suite. I may try to quit kb/mouse software as well, good point
thumb.gif
 
#12 ·
Quote:
Originally Posted by mamuf View Post

15 minute test is definitely not enough for stable OC. It's OK to find the limits. Try running RB for at least 2 hours. I had LuxMark crashing after 20-30 minutes (of course this does not say the CPU is unstable) and RB could BSOD even after few hours.

Also when you get BSOD, what does it say? If you miss the actual BSOD, go into Event Log under Windows Logs -> System and look for errors or even warnings. In my case, unless it's a WHEA error, LuxMark crashed after gpu driver failure which was logged only as a warning.

I am still trying to find out what causes the LuxMark crash. Already tried to close Afterburner (I use it to control gpu fan) and other programs like ASUS AI Suite. I may try to quit kb/mouse software as well, good point
thumb.gif
The 15 minutes test was just to find out if the problem was solved because I would have BSOD or LUXMARK crash already in the first minutes. I did a longer RB test and everything is alright
biggrin.gif


For me closing the background programms I said solved my problem
smile.gif


Good luck!
 
#13 ·
Quote:
Originally Posted by CryWolf View Post

The 15 minutes test was just to find out if the problem was solved because I would have BSOD or LUXMARK crash already in the first minutes. I did a longer RB test and everything is alright
biggrin.gif


For me closing the background programms I said solved my problem
smile.gif


Good luck!
Successfully passed RB 8 hours after closing almost everything I could - those utility apps for kb/mouse, and, of course, AI Suite and other hw/monitoring stuff except for hwinfo. So I think It's done
thumb.gif
 
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top