GPU Problems, please help!

Discussion in 'Videocards - AMD Radeon' started by lenix, Oct 4, 2009.

  1. lenix

    lenix Master Guru

    Messages:
    299
    Likes Received:
    0
    Hello,

    I've been having some trouble with my GPU the past couple of days... At first I used to lose video signal while I'm gaming, it was causing a lot of trouble.
    So I decided to take out everything out of the case and test the system, the problem seemed to go away.. no more crashing while gaming etc.
    But I still face a problem, when I boot the system in the morning, it takes too long for the motherboard to POST ( aka give its first *BEEP* ), takes around 1 minute or so (during the time the monitor is still on stand-by mode), then I'd get kind of weird looking white-dashes all over the screen in bios ( yet i can use the system just fine but it cant enter windows ), after one more reboot everything goes back to normal. This problem doesn't happen on each boot, it sometimes happens, sometimes not.

    It's really confusing, now I don't get any problems while I'm gaming, I did 3dmark06 benchmark and it gave me around 17k points.. I played hours of NFS Shift without any problems.. So I'm not sure what's the faulty hardware here, could it be the motherboard or should I blame the 4870x2? ( I'm pretty experienced with computers but my heart tells me that the video card is fine, i feel something is screwed with the mobo, although all the problems I'm getting is video related )

    Mind you, no heating issues, everything is running cool here, max temps for the 4870x2 are 85C and I keep monitoring the temps the whole time I'm using the pc.

    http://www.dailymotion.com/video/xap3g2_gpu-problem Here is a video describing the problem, sometimes these blue stripes show, other times white dashes show at POST screen.
     
  2. ForgedReality

    ForgedReality Master Guru

    Messages:
    216
    Likes Received:
    0
    It's definitely a hardware problem. Can you try a different power supply? Are you sure the card is pushed all the way into the slot?
     
  3. Pill Monster

    Pill Monster Banned

    Messages:
    25,234
    Likes Received:
    0
    It's most likely the video card judging by the artifacts on your screen.

    Also when your PC bsod, I saw what looked like an ATI.xxxxx file referenced - which also points to your video card..

    Have you overclocked it?
     
    Last edited: Oct 4, 2009
  4. Bodar

    Bodar Ancient Guru

    Messages:
    2,396
    Likes Received:
    0
    It sounds like the card unfortunately.. I dropped a cup of water on a gtx 200 series once and it was'nt happy :)
     

  5. lenix

    lenix Master Guru

    Messages:
    299
    Likes Received:
    0
    Adding to the thread, the problem I talked about ( with the blue stripes etc on boot ) was while the parts were inside the case, I took out the parts again and put them on a wood piece, No problems so far. Been looping Crysis benchmark at highest for 2 hours.. rebooted many times and it seems to function normally..

    Could the case be the reason of the malfunction I used to have? I'm gonna keep the hardware outside the case for 2 days to judge.. but I'd like to know what you guys think.

    Also, nothing is oc'd.
     
  6. Pill Monster

    Pill Monster Banned

    Messages:
    25,234
    Likes Received:
    0
    Well normally in that case I'd would say it could be heat related - but - you were getting artifacts during POST, which normally means the problem is NOT heat related because the chips are still cold.

    Was it cold, or had it already been running for a long time when you made that video?

    If the vram is only partially damaged, then the errors can be intermittent, because they will only show when the bad part of the memory is being addressed.

    If you ran the Crysis benchmark it may have used the same undamaged address space the whole time it was running.

    A good way to tell is to use an app that will use a lot of ram. Actually you can use OCCT to test vram come to think of it...

    Intermittent faults can be a pain to troubleshoot sometimes.
     
  7. lenix

    lenix Master Guru

    Messages:
    299
    Likes Received:
    0
    True to all what you said, the machine was pretty cold when I shoot the video, it was shutdown for like 8 hours in a cold room before I booted it but I was expecting what happened that's why I had the camera recording.

    I'll give OCCT a go, any other intensive applications that can test the whole VRAM is welcome.

    Cheers for the suggestion mate


    Edit: I just installed OCCT, GPU Memtest is disabled because my card doesn't support CUDA, but there is another test available called "GPU:OCCT", will give it a shot, if there any other programs please say :)
     
    Last edited: Oct 4, 2009

Share This Page