Gtx670 crashing

System: gigabyte gtx670, 3570k CPU, 16gb corsair vengeance ram, asus sabertooth mobo, seasonic 750 gold psu, 120 gb samsung ssd and 2.5 tb Seagate hdd, noctua nh d14 in a cooler master 690 case

My first build worked fine for a few weeks until I started to play arkham asylum and about 5 min into the cutscene opening, my fans all turned up to max and the screen went blank. After letting the machine cool down for a couple of min, it boots up fine but won't get past the opening of the game. I figure the problem is with the game and uninstall.
I have been running borderlands 2 on high settings for a couple of weeks with no issues at all and after installing precision x, I up the GPU clock from 1060 to 1100 and run heaven benchmarker at high settings. It gets a couple of minutes in and I get the same noisy crash, so I reset the card to stock and run again, it still crashes. I figure I have a GPU issue and decide not to mess with the settings again.
A couple of days later borderlands 2 crashes the same way an hour into a game and now won't recognise my save file.
Now my new install of Starcraft 2 crashes the exact same way, blank screen, fans all suddenly max out until I hold the power button for 5 sec and I am stuck knowing what to do.

All drivers are up to date for GPU, mobo etc. and when running GPU z or CPU z all my temps seem low.
Is there a way I can diagnose exactly what was happening prior to the crashes? I figure I have a GPU problem but don't know how to find out for sure or what to do about it now? I bought the GPU from scan in the uk but don't have the original box and live in Switzerland so sending it back would be a major hassle, especially seeing as I don't know for sure that it is causing the problem.
I suppose I could buy the same card locally and if everything works ok, send back the original if I am still able? But am hoping to be able to diagnose and fix with the help of some of the experts here.
14 answers Last reply
More about gtx670 crashing
  1. So have you connected both the 6 pin power connectors on the 670?
    What version of windows?

    You will need to check temps while your gaming using the OSD in Precision or Afterbuner. If your problems began with the Precision software I would suggest you uninstall in and try something else like MSI Afterburner.
  2. Yep, GPU has 6 pin and 8 pin and they are both connected. AM running windows 7.
    The first crash was prior to installing precision x, all the rest have been since.
  3. Monitoring with OSD I can see that my temps slowly rise to about 50c just before the crash, never higher. The odd thing is that according to the stats and confirmed by CPU-Z, my clock should be 980 with 1058 boost. But precision x shows it running at 1142 the entire time a cutscene is on. I am not overclocking so I don't know where this number is coming from? Could it be the cause of the crashes? It can't be overheating if I never get I've 50?
  4. download msiafterburner and tell us what temps ur getting.

    when ur ingame.
  5. iceclock said:
    download msiafterburner and tell us what temps ur getting.

    when ur ingame.

    According to Precision X OSD, ingame temps for Borderlands 2 and Starcraft 2 are in the low 40s.
    Cutscenes are high 40s to 50, at which point it crashes.
    voltage is 1.175 the whole time and clock ranges from 980 to 1142 despite the fact the boost clock should only go to 1059
  6. have u modified voltage and clockspeeds on ur card?
  7. iceclock said:
    have u modified voltage and clockspeeds on ur card?

    I used Precision X to move my GPU clock offset up 20 MHz before running Heaven the first ever time, but after that crashed, the machine has had everything set to stock. (ie, Power target 100%, GPU clock offset +0MHz and MEM clock offset +0MHz)
  8. hows it running on stock voltage and speeds?
  9. iceclock said:
    hows it running on stock voltage and speeds?

    Has always been running stock, I only upped the clock for one single run through of heaven then reset to stock.
    Borderlands 2 runs fine with settings turned up high with an in game GPU temp of around 45c.
    In game Starcraft 2 is the same, only problem happens when cutscene comes on. Voltage stays 1.175 and the temp goes to high 40s or 50 and never higher and almost always crashes. I have never got through a heaven benchmark run, regardless of settings without it crashing, again, despite never going over 50c and running 100fps until the crash.

    I'm thinking it has something to do with the fact that it seems to be running at a clock speed higher than my highest boost speed, but I don't know why it is doing that.
  10. have u tried uninstalling ur drivers. than using driversweeper to remove traces of ur old drivers. than reinstall the latest drivers.
  11. iceclock said:
    have u tried uninstalling ur drivers. than using driversweeper to remove traces of ur old drivers. than reinstall the latest drivers.

    Yep, did exactly this a couple of days ago. it made no difference.
  12. weird.
  13. iceclock said:

    So tonight I used Precision X to throttle the card down by 101MHz so it runs at 1019MHz during cutscenes and Heaven benchmark, and I haven't had a crash. It's gone through a couple of full runs at 47, 48C all the way through, which it has never managed before.
    I think I need to get a new card. I guess I should be safe simply replacing this one with the same model as it is clearly a defective model.
  14. u know u can just call up the company u bought the videocard from, the maker of the card, and rma it.

    sounds like its got defects.
Ask a new question

Read More

Graphics Cards Graphics