RE: GPU dying help please

[ed.d.grout]

Hi all,

So recently my MSI Lightning GTX 770 has suddenly started crashing. At first it would just crash the drivers and recover, now (particularly in heavier games) i get a black screen with white square alternating pink all over the screen.

Things I've tried so far:

Messing with the power settings on Nvidia control panel and windows, also disabling vsync

Wiping all drivers and installing as new

Cleaning out all dust from case

It seems to be getting progressively worse, a few minutes ago it crashed on a video and I have to hard reset PC to get any display back. I've switched to my integrated gpu for now while I troubleshoot.

Main question: is this the death of my GPU, it's 4 years old now, is it time to upgrade components?

Cheers for any help,
Ed

 

[jankerson]

System Specs?

 

[ed.d.grout]

Oops apologies, below:

Motherboard: Gigabyte GA-Z87X-UD4H Socket 1150 HDMI DisplayPort DVI ATX Motherboard
GPU: MSI Lightning GTX 770
CPU: Intel Core i5 4670K 3.40GHz Socket 1150 6MB Cache Retail Boxed Processor
RAM: Corsair 8GB (2x4GB) DDR3 1600MHz Vengeance Pro Kit - Blue
PSU: Corsair CX 500W Fully Wired 80+ Bronze Power Supply
SSD: Samsung Pro 500GB
HDD: WD Red 1TB 3.5" SATA NAS Hard Drive
OS: Windows 7 Home Premium

 

[ThatVietGuy]

Symptoms sound reasonable for the GPU failing. If you have another computer to test it in that would be the best confirmation.

 

[jankerson]

Oops apologies, below:

Motherboard: Gigabyte GA-Z87X-UD4H Socket 1150 HDMI DisplayPort DVI ATX Motherboard
GPU: MSI Lightning GTX 770
CPU: Intel Core i5 4670K 3.40GHz Socket 1150 6MB Cache Retail Boxed Processor
RAM: Corsair 8GB (2x4GB) DDR3 1600MHz Vengeance Pro Kit - Blue
PSU: Corsair CX 500W Fully Wired 80+ Bronze Power Supply
SSD: Samsung Pro 500GB
HDD: WD Red 1TB 3.5" SATA NAS Hard Drive
OS: Windows 7 Home Premium

Yeah I would go ahead and replace the Graphics card.

 

[ed.d.grout]

Ok cool thanks, was looking at the 1070. Anything else you would upgrade while I'm at it, might as well do it all at once.

 

[ed.d.grout]

Symptoms sound reasonable for the GPU failing. If you have another computer to test it in that would be the best confirmation.

No other computer unfortunately.

 

[jankerson]

Ok cool thanks, was looking at the 1070. Anything else you would upgrade while I'm at it, might as well do it all at once.

Yeah, replace the PSU also.

The CX500 isn't really a good one.

Look at the Seasonic Focus 550W or 650W.

 

[RaDiKaL_]

Just to cover everything make sure to check its temps, could be simply time to renew its thermal paste or its fans are failing.

+1 to jankerson's comment, Corsair CX series have been known to decay after a couple years of use (some even after only 6 months) leading to issues like yours.

 

[ed.d.grout]

Just to cover everything make sure to check its temps, could be simply time to renew its thermal paste or its fans are failing.

+1 to jankerson's comment, Corsair CX series have been known to decay after a couple years of use (some even after only 6 months) leading to issues like yours.

I checked temperatures and it was running very cool, 35-40 degrees. Can't test at full load as it crashes on anything that puts any load on it. I tried putting fans at 70+ % and then running and still the crashing. It's got so bad now, even when I load up battle net it will crash instantly and say Kernel Driver crashed and recovered successfully.

 

[RaDiKaL_]

Use HWInfo and enable its logging feature, this should provide us info of the max temp reached and more importantly how your voltages are right before the crash occurs. After getting back into windows from the crash look for the log file and review it or upload it if you cannot read it so we can check it.

 

[ed.d.grout]

Use HWInfo and enable its logging feature, this should provide us info of the max temp reached and more importantly how your voltages are right before the crash occurs. After getting back into windows from the crash look for the log file and review it or upload it if you cannot read it so we can check it.

Ok logged the crash, saved below:
https://filebin.net/y1dor7em8ibhf6ks

Doesn't mean anything to me, not even sure what to look for


Should mention, it crashed once driver only. I then opened up another game and after a while it crashed completely and I had to hard restart.

 

[ed.d.grout]

Another file log of multiple driver crashes and a final crash with hard reset at the end:
https://filebin.net/x08zbkv1ip7m1g2j

Also image of what happens when both crashes occur:
https://imagebin.ca/v/3dt16Nl5giiV

 

[RaDiKaL_]

Ok there are two issues:

- Your CPU is reaching 95ºC+ so you need to fix that immediately.

- For some reason your GPU pulls no more than 23W even at full usage, without enough power its obvious that it's gonna crash since it still reaches up to 80% usage though its core clock stays at 540mhz (I don't know what is the max freq for the GTX 770).

Your PSU voltages seem fine, so maybe you didn't plug all the PCI-E power connectors correctly into your GPU? double check that to see if the problem disappears.

 

[loslyfgmail]

I've been having serious problems with random BSOD with my MSI GTX 770 and the latest Nvidia drivers. I rolled back to Nvidia driver version 382.05 - 05/04/2017 and my system has been rock solid again for over 3 weeks now. Please try the old driver and let us know.
http://www.tomshardware.co.uk/answers/id-3529154/random-bsod-latest-nvidia-drivers.html

 

[ed.d.grout]

Ok there are two issues:

- Your CPU is reaching 95ºC+ so you need to fix that immediately.

- For some reason your GPU pulls no more than 23W even at full usage, without enough power its obvious that it's gonna crash since it still reaches up to 80% usage though its core clock stays at 540mhz (I don't know what is the max freq for the GTX 770).

Your PSU voltages seem fine, so maybe you didn't plug all the PCI-E power connectors correctly into your GPU? double check that to see if the problem disappears.

Apologies, work long weeks only had a chance to have a look at it again now. Manged to fix the CPU heat, was a bunch of dust that had got trapped between the heatsink and the CPU. It's now running at 49C.

GPU still bad, will try your solution now loslyfgmail, thanks for letting me know

 

[loslyfgmail]

So, a slight update - I just recently did the Windows 10 Fall Creators update, and during that my Nvidia driver was automatically updated to 385.xx and once again I started suffering from black screen of death at random times. I decided to try the latest driver 387.92 date 9.10.2017 and to my surprise it worked! So, I'd first recommend an old stable driver like 382.05 - 05/04/2017 for the GTX 770. If that works then you know it's driver related, then try the latest 387.92. Good luck and let us know if that worked for you.

 

[ed.d.grout]

I've been having serious problems with random BSOD with my MSI GTX 770 and the latest Nvidia drivers. I rolled back to Nvidia driver version 382.05 - 05/04/2017 and my system has been rock solid again for over 3 weeks now. Please try the old driver and let us know.
http://www.tomshardware.co.uk/answers/id-3529154/random-bsod-latest-nvidia-drivers.html

Unfortunately tried 6 different drivers using ddu to roll back, and none of them worked. It's strange, I'm receiving zero performance issues, just kernel driver crashing over and over until suddenly I get the screen above permanently with crazy noises coming from the speakers and the computer either turns itself off or just stays like it.

I can't afford a PSu as of yet to see if that May be causing the issue. Any other suggestions are welcomed in the meantime.

 

[loslyfgmail]

Sorry to hear that. The only thing I'd recommend is try the oldest driver you can find. If it still crashes then it is probably hardware related. Also, try disabling/removing your graphics card, if the problem stops then at least you're sure that's what it was. Unfortunately, you'll then just have to bite the bullet and get another graphics card. The GTX 1070 is looking very desirable for the price...

 

[ed.d.grout]

Sorry to Necro a thread, just hate finding posts with no follow-up. Essentially nothing work, card was busted, tried new PSU etc. Am rebuilding new PC now.