Is my GPU slowly dying?

[AGGELOS_970]

Hello guys!

 

So I have a GTX 660Ti from MSI. Two times one of the two fans stopped working. 

First time I replaced them, second time I didn't bother. I just removed my overclock.

So for a lot of months now my GPU is running with only one of two fans. And depending on the game, if it's maxing out my GPU it is running on about 88 degrees celsius (190 Fahrenheit).

I didn't have any issues though. So far...

 

Today I was playing Rainbow Six Siege and suddenly my game froze, and after a while my PC restarted on its own.

Went into Windows and after 5 minutes everything froze except the mouse pointer. 

Restarted the PC manually. This time no issues on Windows.

Started the game again and after some time it froze again. It didn't restart on its own this time. Task manager would not open so I just restarted it manually.

Tried another game and again it froze the whole system. After restarting my PC I had no signal from the GPU on my monitor. 

I unpluged the HDMI cable and I pluged it on my motherboards port. So I was using Intel graphics.

I downloaded DDU and reinstalled my graphics drivers.

This time I had signal coming ouy of my GPU. Tried playing some games but again the PC just froze again after a while.

Restarted with no signal coming from the GPU again. After a 5 minutes wait and a shutdown in between I had signal on my GPU. 

So now I'm using my GPU. But I won't try playing any games again cause I'm certain it will crash again.

 

I'm pretty sure that's the GPU slowly dying cause of running too hot for some time now.

Like, nvidia says my max GPU temp should be 90 and I was running ~88 on some game sessions. 

 

But yesterday I was playing fine. 

Also I am using RTSS for capping my framerate now. So my GPU isn't fully stressed on all games (some do stress it though).

Like the last time that it crashed I think it was running on 70 celsius. 

 

So that's why i'm not 100% sure. Could it be another part that is failing? 

I don't think it's a software issue because of those two occasions that my GPU had no signal after the crashes. 

 

What do you guys think?

I am already planning on getting a GTX 1060 pretty soon so that's good. 

But I want to be sure so I don't get the new GPU and have the same problems again.

 

Thanks!

[AGGELOS_970]

1 minute ago, wrathoftheturkey said:

Your GPU stopped working a while ago and now you're asking if your GPU is broken? I'd bet my left pinky on it (so confident, but not that confident)

Yes but now Im on my GPU again

[Cereal5]

try uninstalling your drivers with DDU and reinstalling the newest one from the NVIDIA website. If still no, then try unplugging your GPU and using onboard graphics if your motherboard supports it, and if everything works well, its  safe to assume it is

[AGGELOS_970]

2 minutes ago, Cereal5 said:

try uninstalling your drivers with DDU and reinstalling the newest one from the NVIDIA website. If still no, then try unplugging your GPU and using onboard graphics if your motherboard supports it, and if everything works well, its  safe to assume it is

Yes I did the DDU thing already. 

Im on my GPU now and its fine on the desktop but it will pretty sure crash in game again. 

I will try running a GPU stress test and see if I get anything from there.

[Edgar R. Zakarian]

5 hours ago, AGGELOS_970 said:

Hello guys!

 

So I have a GTX 660Ti from MSI. Two times one of the two fans stopped working. 

First time I replaced them, second time I didn't bother. I just removed my overclock.

So for a lot of months now my GPU is running with only one of two fans. And depending on the game, if it's maxing out my GPU it is running on about 88 degrees celsius (190 Fahrenheit).

I didn't have any issues though. So far...

 

Today I was playing Rainbow Six Siege and suddenly my game froze, and after a while my PC restarted on its own.

Went into Windows and after 5 minutes everything froze except the mouse pointer. 

Restarted the PC manually. This time no issues on Windows.

Started the game again and after some time it froze again. It didn't restart on its own this time. Task manager would not open so I just restarted it manually.

Tried another game and again it froze the whole system. After restarting my PC I had no signal from the GPU on my monitor. 

I unpluged the HDMI cable and I pluged it on my motherboards port. So I was using Intel graphics.

I downloaded DDU and reinstalled my graphics drivers.

This time I had signal coming ouy of my GPU. Tried playing some games but again the PC just froze again after a while.

Restarted with no signal coming from the GPU again. After a 5 minutes wait and a shutdown in between I had signal on my GPU. 

So now I'm using my GPU. But I won't try playing any games again cause I'm certain it will crash again.

 

I'm pretty sure that's the GPU slowly dying cause of running too hot for some time now.

Like, nvidia says my max GPU temp should be 90 and I was running ~88 on some game sessions. 

 

But yesterday I was playing fine. 

Also I am using RTSS for capping my framerate now. So my GPU isn't fully stressed on all games (some do stress it though).

Like the last time that it crashed I think it was running on 70 celsius. 

 

So that's why i'm not 100% sure. Could it be another part that is failing? 

I don't think it's a software issue because of those two occasions that my GPU had no signal after the crashes. 

 

What do you guys think?

I am already planning on getting a GTX 1060 pretty soon so that's good. 

But I want to be sure so I don't get the new GPU and have the same problems again.

 

Thanks!

My dad was playing that rainbow six as well, his PC restarts in the middle of the game as well. All other games run fine on his 760 gtx from evga.

[jools]

gpu-z, furmark

mite be worth reseating heatsinks and get that fan working or some blower on gpu heatsink

88c is warm to hot 90c being hot

also get latest drivers

should get enuff life outta card till 1060 arrives

[TheRandomness]

Right, so you're saying there could be a temperature issue but you keep using the graphics card anyway? I'd say remount the cooler. I've had similar issues with my 290 when it got too hot. 

[AGGELOS_970]

5 hours ago, rrubberr said:

I have a couple 660ti's that ran a lot hotter than that for their whole lives (especially the bottom one in SLI), so I might check for drivers and do a fresh install of Windows before making any assumptions.

 

5 hours ago, TheRandomness said:

Right, so you're saying there could be a temperature issue but you keep using the graphics card anyway? I'd say remount the cooler. I've had similar issues with my 290 when it got too hot. 

 

5 hours ago, jools said:

gpu-z, furmark

mite be worth reseating heatsinks and get that fan working or some blower on gpu heatsink

88c is warm to hot 90c being hot

also get latest drivers

should get enuff life outta card till 1060 arrives

 

11 hours ago, Cereal5 said:

try uninstalling your drivers with DDU and reinstalling the newest one from the NVIDIA website. If still no, then try unplugging your GPU and using onboard graphics if your motherboard supports it, and if everything works well, its  safe to assume it is

 

11 hours ago, wrathoftheturkey said:

Your GPU stopped working a while ago and now you're asking if your GPU is broken? I'd bet my left pinky on it (so confident, but not that confident)

 

 

Ok so I'm running stress test for the past 2 hours now and got nothing.

No crashes no anything. 

The fist half an hour I was running Aida64 along with Unigine Heaven and after that it was Aida64 with FurMark.

I also checked out the boxes for: Stress local disks, system memory, cache and FDU.

So the system had everything working maxed out.

I also did 2 runs with Memtest86+ and found no errors.

What do you make out of this?

 

I will try playing some games again and see what happens and I'll post back when I got any news.

 

Here some screenshots you can check:

[jools]

3 hours ago, wrathoftheturkey said:

Reinstall the programs, reinstall Windows. That's weird af

yep fresh OS ,weve stressed hardware now new software

[AGGELOS_970]

3 hours ago, wrathoftheturkey said:

Reinstall the programs, reinstall Windows. That's weird af

I played Siege for around 3 hours and I had no issues.

I didn't change anything from yesterday.

I dont know whats going on but as you said that weird af lol.

Like... how can it be a software issue?

But then again how can it be a hardware one?

Stress test for 2 hours and nothing happened.

 

I will continue using my PC like I was used to and I will try out some more games.

If anything new comes up I'll post here.

[AGGELOS_970]

3 hours ago, wrathoftheturkey said:

Reinstall the programs, reinstall Windows. That's weird af

 

3 minutes ago, jools said:

yep fresh OS ,weve stressed hardware now new software

 

Yeah if it happens again I'll reinstall Windows.

[jools]

future ref... use cpu-z with prime95  and gpu-z with furmark for testing at full screen resolution, these give best full load results. obviously you test cpu first with (cpu-z with prime95) then stop cpu tests and run gpu (gpu-z with furmark), not together(cpu+gpu tests).

[AGGELOS_970]

8 minutes ago, jools said:

future ref... use cpu-z with prime95  and gpu-z with furmark for testing at full screen resolution, these give best full load results. obviously you test cpu first with (cpu-z with prime95) then stop cpu tests and run gpu (gpu-z with furmark), not together(cpu+gpu tests).

Ok thanks if it happens again I'll do it that way.

What's the use of cpu-z and gpu-z tho? 

Only for monitoring?

[jools]

yep better monitoring same as aida but most techys use cpu-z and gpu-z , download them have a look. I just think they are a bit easier to read data from. they are free. plus you concentrate on one bit of hardware a time instead of getting everything but your call as always