1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

Card Crashing - GTX 770

Discussion in 'Graphics Cards' started by Longbow, Mar 21, 2015.

  1. Longbow

    Sgarrista

    Joined: Jan 15, 2004

    Posts: 9,339

    The card is a KFA2 GTX 770, when I play games, especially ones heavy in graphics, the card seems to stop working. I'll see the game freeze for a split second, then hear the fans stop on the card, before it kicks back out to desktop.

    I thought initially it was a driver conflict, but I have since then installed a fresh OS.

    [​IMG]

    You can see the large drop just after the dotted line. Card was at 80c.

    I've taken the card out and put in back in several times, so it's unlikely to be that.

    Power supply is Corsair HX520W which should be sufficient?

    Any ideas?
     
  2. PieEater

    Mobster

    Joined: Oct 18, 2002

    Posts: 3,507

    Location: Devon

    Could possibly be temperature related, when you took the card out did you check the cooler to make sure the fins were free of dust and the fans spun freely?
     
  3. DavyBoy

    Mobster

    Joined: Dec 6, 2008

    Posts: 2,693

    Location: Burghead, Elgin

    Do you have any error reports/warnings in your Event Viewer?
     
  4. J.D

    Mobster

    Joined: Jul 26, 2006

    Posts: 4,788

    Location: Edinburgh

    I doubt it's heat related but try playing with your fans at 100%. If the temp drops from 80'C and your game works as intended then it's heat related.

    Saying that, I'd try different drivers tbh. I've had one driver that crashed BF4 to desktop four times in three matches. Reset my motherboard BIOS, overclocked my CPU from scratch again and the problem remained until I went back to an older driver. Also from reading recently, the drivers have been causing headaches for some so at least try a few different sets when testing.
     
  5. Longbow

    Sgarrista

    Joined: Jan 15, 2004

    Posts: 9,339

    I took out the card, got rid of what little dust there was, I didn't run the fans at 100% but I kept the case cover off and put a big fan next to it, but it still crashed, even though it was still hitting 80c, but when it crashed, it was at 71c:

    [​IMG]

    Edit: Tried it again but forced the fans at 85%, still crashed at 75c.

    It's not just CryEngine related as it happens with other games.

    "Display driver nvlddmkm stopped responding and has successfully recovered."

    Well I installed a newer driver the other day and it doesn't seem to change it, shall I just go back several drivers?
     
    Last edited: Mar 22, 2015
  6. PieEater

    Mobster

    Joined: Oct 18, 2002

    Posts: 3,507

    Location: Devon

    I could be wrong but it still looks temperature related to me as if your VRMs are overheating and throttling the power to the GPU which then becomes unresponsive causing the crash. Have you changed your cooler at all? Try running with the fans at 100% to see if this has any effect the additional air movement should cool the VRMs a bit more.

    /Edit:- Can you change your Afterburner skin to one that shows your power limit and temp limit settings and confirm what these are set to.
     
    Last edited: Mar 22, 2015
  7. Divine Pain

    Hitman

    Joined: Jan 9, 2012

    Posts: 907

    Location: Wirral

    That error is driver related.

    Do a repair/reinstall of directx and vcredist drivers.

    And do a clean uninstall/reinstall of the nvidia drivers (make sure to check: "clean installation")

    Try that see if that changes anything.

    EDIT: Dont forget to restart!
     
    Last edited: Mar 22, 2015
  8. Longbow

    Sgarrista

    Joined: Jan 15, 2004

    Posts: 9,339

    I completely formatted and it's still doing it.

    I ran the fans at 100%, and added GPU max temp to the graph. It seems 80c is the max so it was constantly hitting this.

    I used Nvidia Inspector to set a frame rate limiter at 75c. It crashed at 78c.

    [​IMG]
     
    Last edited: Mar 22, 2015
  9. PieEater

    Mobster

    Joined: Oct 18, 2002

    Posts: 3,507

    Location: Devon

    [​IMG]
    The above was taken for another thread so ignore everything else apart from the AfterBurner skin, you can see that there are settings for power limit and temperature limit, these do not show on the default MSI skin that you are using so if you go to settings and user interface have a look for one of the newer skins that has this info (I'm assuming it should be there for the 770). It would appear that your temperature limit is set to around 80c at which point your card is throttling which may possibly be the reason behind your crashes if the 3d app is not getting sufficient frames. The maximum safe temperature for your card is 98c so you could safely raise the temp limit to 90c making sure your fans are at 100% at these temps, see if this has any impact, if not we know for sure I'm barking up the wrong tree. Whilst you're there also check your power limit slider as this appears to be capped also unless it is throttling due to heat.
     
  10. DavyBoy

    Mobster

    Joined: Dec 6, 2008

    Posts: 2,693

    Location: Burghead, Elgin

    That is the exact same error that I am getting with my GTX 970, I've tried absolutely everything that I can think of to sort it, but I am thinking that maybe I've been unlucky and received a bad card.

    I've put my BIOS right back to defaults just to see whether it might have been my overclocked settings that were the cause, but no, it still gives me that same error, yet I can pop in an ancient GTX 470 and my system is perfectly stable, although it really struggles with pretty much every new game out there it still runs them on lower settings without this crash ever happening.

    God knows what it is that's causing it for me, it's certainly not temps as it has never gone above 57c when running 3d mark or gaming.

    I'm going to test it with a different PSU, and I will also test it out on another PC as well, it's really annoying because the graphics in my games now look really good and it runs smooth too, I just can't stop this error from happening regardless of what I've tried.

    The Microsoft Fix It for TdR settings causes it to crash to a grey screen, I've underclocked the GPU and the vRAM, BIOS settings at defaults, changing Power Management Mode in the nVidia control panel from Adaptive to Prefer Maximum Performance, but none of this has made a blind bit of difference for me.
     
  11. Longbow

    Sgarrista

    Joined: Jan 15, 2004

    Posts: 9,339

    I didn't know about that.

    I changed it and upped the temp max to 94c. However it still stops, last attempt at 85c, previous attempts at 88c and 90c, max was always set to 94c.

    [​IMG]

    Unless you can see I'm doing something wrong?

    Let me know how that goes! Unfortunately I don't have another PSU laying around. It sure isn't exceeding the wattage of the PSU though.
     
  12. PieEater

    Mobster

    Joined: Oct 18, 2002

    Posts: 3,507

    Location: Devon

    What I'm seeing throughout this is that your cards GPU temperature fairly quickly climbs to it's max allowed setting, at which point it will throttle which then appears to cause it to crash. I'm also seeing that your GPU power never gets above 85% when it should be at 100% most of the time only reducing in the event of overheating. So I recon you have an overheating problem, which you may be able to get round by taking your cooler off, giving it a damn good clean and refitting with some quality thermal paste (but not if your card is still under warranty as this may void it). However the overheating could be caused by a fault with the card such as a faulty VRM allowing too much power through, in which case you may be looking at repair / replacement.
     
    Last edited: Mar 22, 2015
  13. kaku

    Wise Guy

    Joined: May 21, 2013

    Posts: 1,800

    Maybe take the side off your PC and bang a desk fan pointing at the card just to pinpoint if it's a heat issue.

    I would probably say it's heat or VRM/power issue at this point.
     
  14. DavyBoy

    Mobster

    Joined: Dec 6, 2008

    Posts: 2,693

    Location: Burghead, Elgin

    Hey man, it turned out that it was actually my custom braided 24 pin cable extension that was messing my system up.

    My dad took down his PSU, and as I was disconnecting my one I had a look inside the socket of the extension cable and noticed that several of the pins were pushed way back in the socket, I'd noticed in the Voltage Monitoring section in my BIOS that the +12v rail was sitting at 13.550v and the 3.3v rail was highlighted in red and was down at 2.8v, anyway as soon as I connected the PSU without the extensions and I entered my BIOS to look at the voltages they were all spot on, so I booted into Windows played games the entire night and did several runs of 3d Mark, then left it running Unigine Heaven the whole night and it is running perfectly now, no more irritating driver crashes.

    You had any success with your problem?
     
  15. Longbow

    Sgarrista

    Joined: Jan 15, 2004

    Posts: 9,339

    An update:

    [​IMG]

    Removed the TIM and re-applied some more.

    Temperatures are slightly better, but it was still crashing.

    I've set the temp limit lower to 75c and that seems to work. What is weird is before when the temp limit was higher at say 90c, it would sometimes crash at 70c, yet with the limit at 75c, it will happily sit there.
     
  16. Sean_UK

    Mobster

    Joined: Jul 6, 2006

    Posts: 2,507

    Location: South East

    Up the voltage via afterburner .13v or .25v if you can? That or drop the core clock by 50 and see if crashing stops.