Advertisement
If you have a new account but are having problems posting or verifying your account, please email us on hello@boards.ie for help. Thanks :)
Hello all! Please ensure that you are posting a new thread or question in the appropriate forum. The Feedback forum is overwhelmed with questions that are having to be moved elsewhere. If you need help to verify your account contact hello@boards.ie

Something fishy in my machine.

  • 19-12-2013 11:23am
    #1
    Registered Users, Registered Users 2 Posts: 371 ✭✭


    So, like quite a few people, I've run into the grey screen of death, crashing to grey vertical lines on the screen, with an AMD card. It's a HD 6950, 2gb.

    I've really gone round the houses on it, did all the wiping drivers, re-installing them at different versions and so on, re-seating the card, etc, etc.

    The only thing that's seemed to make a difference so far is underclocking the gpu. I dropped it from the default 800mhz to 750mhz. It seems to be a lot more stable now, but it's not as if I'd overclocked it before that. I also increased the power to add 15%, which didn't do anything before, but I left it like that.

    For example, I started running the DayZ alpha yesterday. Kept getting the grey screen of death within five minutes. After underclocking, it ran quite happily for 25 minutes, at which time I called it a night.

    The previous thing I'd tried, which is still in place, was changing registry settings so Windows would leave more time before restarting the driver when it didn't get a response from the graphics card.

    Temperatures have always been fine.

    There's a chance the card is just on it's way out, but there's one thign that's occurred to me this morning. I've a fan speed controller for the case fans, but it hasn't been working for months. I've never bothered about it because my temperatures have been ok. Could there be a connection?

    Maybe I should be looking at the power supply or the motherboard, and not the graphics card? What would you look at?

    Mobo: ASRock P67 Pro3 (B3), Sockel 1155, ATX
    CPU: Intel Core i5-2500K
    Graphics Card: Sapphire HD 6950 Dirt3, 2GB GDDR5 PCI-Express
    PSU: Super-Flower Amazon 80Plus 550W


Comments

  • Registered Users, Registered Users 2 Posts: 878 ✭✭✭Luck100


    I would double check the power connections to the video card - I think that card has two 4-pin connectors? Make sure they're well seated at both ends.

    Your fan speed controller is dead - so those case fans are running 100% now, or not at all? Either way, I'd fire up precision X or whatever you use for monitoring/overclocking the GPU and see what the temperatures are like. Try running Unigine Valley benchmark with the monitoring on and see what your temps/power draw/voltages/clocks are like under load.

    Even though you're not overclocking, it might be worth bumping the GPU clock up in steps to see where you lose stability when running Unigine Valley. The card just might be on the edge of stability even at stock, in which case you should see instability very quickly if you notch up the GPU clock at all. Same applies to GPU memory clock.


  • Registered Users, Registered Users 2 Posts: 371 ✭✭Fussgangerzone


    Luck100 wrote: »
    Your fan speed controller is dead - so those case fans are running 100% now, or not at all?
    Somewhere around 40%, I'd estimate.
    Either way, I'd fire up precision X or whatever you use for monitoring/overclocking the GPU and see what the temperatures are like. Try running Unigine Valley benchmark with the monitoring on and see what your temps/power draw/voltages/clocks are like under load.
    Would HWMonitor be good enough? I have Catalyst Control Center too, but it gives scant information.
    Even though you're not overclocking, it might be worth bumping the GPU clock up in steps to see where you lose stability when running Unigine Valley. The card just might be on the edge of stability even at stock, in which case you should see instability very quickly if you notch up the GPU clock at all. Same applies to GPU memory clock.
    That sounds like a good idea.
    Will try these tomorrow, thanks.


  • Registered Users, Registered Users 2 Posts: 878 ✭✭✭Luck100


    Would HWMonitor be good enough? I have Catalyst Control Center too, but it gives scant information.

    I think HWMonitor will only show temperatures, but it's better than nothing. You can also try HWInfo32 or HWInfo64 (free), they definitely report much more information than HWInfo.

    In case you do use Unigine Valley, just be aware that the GPU clock speed and memory speed numbers that is displays are wrong (at least on NVidia cards). The other numbers are correct.


  • Registered Users, Registered Users 2 Posts: 878 ✭✭✭Luck100


    Actually, why not try MSI Afterburner? That works with ATI cards. Would be much more useful for GPU testing than HWInfo or HWMonitor.


Advertisement