Is my 1080 Ti at the end of it's life?

[Votivee]

Specs:

5820k

MSI X99S SLI Krait Edition

32GB Corsair Vengance

Corsair RM750

EVGA 1080 Ti FTW3

 

Recently, my GPU had been crashing on specific games after playing for a while. No artifacts, just crashing.

This past week, I've been getting grey screens and crashing in nearly every single game. 

 

I thought it was a temperature issue because the time to the crash gets shorter and shorter after every game launch, but after monitoring temps I'm not seeing any issues. 

 

For example, I just ran a few timespy stress tests. The first lasted maybe 2 minutes before crashing. My most recent stress test lasted not even 40 seconds before crashing. 

 

Anyone have any ideas on what to try next?

 

[jaslion]

Keep in mind the gpu does lower it's clocks dynamically to prevent heat issues but it seems this is fine. Maybe you do wanna do a cleanout. Also maybe remove the oc of the card as it's boosting 300mhz higher than the normal max boost clock from nvidia (sure the ftw version does go higher but this is a lot)

[TheBean]

is your cpu overlclocked? is the gpu overclocked?

try DDU and installing fresh drivers

update windows

 

[Votivee]

Just now, The_Vaccine said:

is your cpu overlclocked? is the gpu overclocked?

try DDU and installing fresh drivers

update windows

 

CPU is overclocked but I haven't changed the OC for years. 

GPU is at factory settings.

I'm on a 100% fresh windows install with fresh NVIDIA drivers

[JaydenTechTips]

Just now, Votivee said:

CPU is overclocked but I haven't changed the OC for years. 

GPU is at factory settings.

I'm on a 100% fresh windows install with fresh NVIDIA drivers

Ok try making a nice profile over clock 1080s are still good and still will be maybe change the CPU's  clock as well.

accommodate the fact there maybe a high bottleneck so fix your overclocks for GPU and CPU. and accommodate the overclock for your cooling solution

[Vishera]

Try testing the GPU with GPUPI 3.3.3,Press Calculate then use the following settings:

32B HWbot GPU

choose your GPU in OpenCL GPU devices

Batch size: 20M

Reduction size: 512

Then press OK to start.

(Press cancel if a message about HPET appears)

 

Tell us if you get any errors while testing.

 

GPUPI 3.3.3 download page:

https://www.overclockers.at/news/gpupi-international-support-thread

[JaydenTechTips]

Just now, Vishera said:

Try testing the GPU with GPUPI 3.3.3,Press Calculate then use the following settings:

32B HWbot GPU

choose your GPU in OpenCL GPU devices

Batch size: 20M

Reduction size: 512

Then press OK to start.

(Press cancel if a message about HPET appears)

 

Tell us if you get any errors while testing.

 

GPUPI 3.3.3 download page:

https://www.overclockers.at/news/gpupi-international-support-thread

Good idea maybe but his or her graphic card is fine

[Votivee]

12 minutes ago, jaslion said:

Keep in mind the gpu does lower it's clocks dynamically to prevent heat issues but it seems this is fine. Maybe you do wanna do a cleanout. Also maybe remove the oc of the card as it's boosting 300mhz higher than the normal max boost clock from nvidia (sure the ftw version does go higher but this is a lot)

This might be going somewhere. 

 

Dropped the clock by 200 (as much as PrecisionX allows) to 1683.

 

I'm 5+ minutes into a timespy stress test now which is farther than I've been getting.

[Vishera]

Just now, JaydenTechTips said:

Good idea maybe but his or her graphic card is fine

If it's fine it will pass the test.

There are tests for RAM and CPU as well in case the GPU pass.

[jaslion]

3 minutes ago, Votivee said:

This might be going somewhere. 

 

Dropped the clock by 200 (as much as PrecisionX allows) to 1683.

 

I'm 5+ minutes into a timespy stress test now which is farther than I've been getting.

Yeah it probably is just having issues with the oc then not a problem.

[Votivee]

7 minutes ago, JaydenTechTips said:

Ok try making a nice profile over clock 1080s are still good and still will be maybe change the CPU's  clock as well.

accommodate the fact there maybe a high bottleneck so fix your overclocks for GPU and CPU. and accommodate the overclock for your cooling solution

I'm having trouble understanding what you're saying here. 

 

It could be a CPU issue, but my OC has been rock solid for 5 years and I trust that it's fine. I don't have any issues when taxing just the CPU.

 

I'm pretty sure there's not bottleneck issue here, my CPU sits around 50-60% most games while my GPU is at 100%.

 

Temps are fine.

[Votivee]

2 minutes ago, jaslion said:

Yeah it probably is just having issues with the oc then not a problem.

Strange that the factory OC would go bad like that. I've had this card for roughly a year and am just now having these issues. 

 

I lowered the clocks by 100 last night but was still having issues. Also tried raising the power limit and it just crashed faster. As of now though we're still going strong in 3dmark timespy. 

 

I'll try some games once this is over but as of now I think your solution worked.

[jaslion]

1 minute ago, Votivee said:

Strange that the factory OC would go bad like that. I've had this card for roughly a year and am just now having these issues. 

 

I lowered the clocks by 100 last night but was still having issues. Also tried raising the power limit and it just crashed faster. As of now though we're still going strong in 3dmark timespy. 

 

I'll try some games once this is over but as of now I think your solution worked.

You may wanna check with msi afterburner and gpu z if there is still good temps on the memory and such.

[Votivee]

Just now, jaslion said:

You may wanna check with msi afterburner and gpu z if there is still good temps on the memory and such.

I've been watching them with PrecisionX and everything is in check.

 

Successful timespy pass:

[jaslion]

1 minute ago, Votivee said:

I've been watching them with PrecisionX and everything is in check.

 

Successful timespy pass:

Yeah probably just age wearing the components down.

[TheBean]

your cpu overclock is probably just unstable now. try going stock settings and check if it still crashes. 

 

Edit: as time passes, the silicon degrades (faster when overclocked) so your overclock will require higher and higher voltage as time passes. 

 

Either reset the overclock or increase cpu core voltage. 

Edited December 13, 2020 by The_Vaccine

[Falkentyne]

You could try a complete thermal re-pad of the GPU and repaste (at the same time).

 

You would have to know the sizes of the stock pads and their locations.  If you can't measure them or aren't sure, then it's usually a safe bet to try compressible 1.5mm thermal pads, as they can work in locations where 1mm pads were previously, by compressing.  Then if there were any 2mm thermal pads (that were not SUPER compressed, meaning the stock pads were a bit too thick), you can always stack 0.5mm pads on top.

 

I suggest Thermalright Odyssey 1.5mm pads (and a second order of 0.5mm Odyssey pads).  They are very good pads.  And for long lasting VERY good thermal paste, Thermalright TFX.  But don't even attempt to spread it.  Just apply it in an "X" pattern corner to corner on the GPU in two large thin (not TOO thin!) diagonal lines and you're good to go.

 

Unfortunately, the best way to get a good price for these pads are from Yao Yue store on Aliexpress, which seems to be an official supplier for Thermalright in China.  You can get a giant 120*120mm 0.5mm pad for $12 and a 1.5mm pad for $18.  While the only pads available on Amazon are the 85mm * 45mm size pads, which are less than half the size of the full 120mm*120mm one

 

Another method of applying TFX is to do 5 medium sized drops, one in the very middle like oldschool, and four in each corner.  That also works.

 

Multiple people have restored their cards to good working order with a proper full thermal pad and thermal paste rework.

Another VERY helpful thing to do is to buy Deoxit D5 and spray it on the power connectors, and into the PCIE slot (obviously with the power supply unplugged).  That has also helped multiple people dealing with contact issues.

[Votivee]

4 hours ago, The_Vaccine said:

your cpu overclock is probably just unstable now. try going stock settings and check if it still crashes. 

 

Edit: as time passes, the silicon degrades (faster when overclocked) so your overclock will require higher and higher voltage as time passes. 

 

Either reset the overclock or increase cpu core voltage. 

Turned out to be the GPU clock. Just played 5 hours of Siege with no issues whatsoever

[Votivee]

49 minutes ago, Falkentyne said:

You could try a complete thermal re-pad of the GPU and repaste (at the same time).

 

You would have to know the sizes of the stock pads and their locations.  If you can't measure them or aren't sure, then it's usually a safe bet to try compressible 1.5mm thermal pads, as they can work in locations where 1mm pads were previously, by compressing.  Then if there were any 2mm thermal pads (that were not SUPER compressed, meaning the stock pads were a bit too thick), you can always stack 0.5mm pads on top.

 

I suggest Thermalright Odyssey 1.5mm pads (and a second order of 0.5mm Odyssey pads).  They are very good pads.  And for long lasting VERY good thermal paste, Thermalright TFX.  But don't even attempt to spread it.  Just apply it in an "X" pattern corner to corner on the GPU in two large thin (not TOO thin!) diagonal lines and you're good to go.

 

Unfortunately, the best way to get a good price for these pads are from Yao Yue store on Aliexpress, which seems to be an official supplier for Thermalright in China.  You can get a giant 120*120mm 0.5mm pad for $12 and a 1.5mm pad for $18.  While the only pads available on Amazon are the 85mm * 45mm size pads, which are less than half the size of the full 120mm*120mm one

 

Another method of applying TFX is to do 5 medium sized drops, one in the very middle like oldschool, and four in each corner.  That also works.

 

Multiple people have restored their cards to good working order with a proper full thermal pad and thermal paste rework.

Another VERY helpful thing to do is to buy Deoxit D5 and spray it on the power connectors, and into the PCIE slot (obviously with the power supply unplugged).  That has also helped multiple people dealing with contact issues.

Thermals are not and never were the issue. 

 

It's been repasted with MX-4 a few months ago. The card happily sits at 83-84c with the fan curve set to silent and gets down around 75c with a more aggressive setting.

 

With the EVGA ICX sensors I can monitor the temps of pretty much everything on the card. Everything besides the GPU stays under 80c