Is the GPU or the PSU failing?

Discussion in '3D Hardware, Software & Output Devices' started by Quitch, Aug 15, 2009.

  1. Quitch

    Veteran

    Joined:
    Jun 11, 2003
    Messages:
    1,521
    Likes Received:
    4
    Location:
    UK
    I've had an issue start recently where my PC will freeze up in games. It will do one of three things:

    1. The image will freeze on the screen while audio continues to play (but the game itself stops completely). All input devices fail to respond. I have to reset the computer.

    2. The screen goes into power saving mode with an "Out of range" error while audio continues to play. I have to reset the computer.

    3. Very occassionally I'll crash to desktop and get a video driver recovery message from the notification area. Sometimes shortly thereafter I'll get a BSOD.

    After a reset Windows will usually load, though now and then I will get a BSOD on boot with an error that the video card was reset and the timeout period expired and the computer only works after another reboot or two.

    This issue has been occurring with Fallout 3, though whether that's relating to something in the game or simply that I've played it a lot recently I don't know. Usually it'll happen after a long period of playing, but once or twice it happened as the opening menu loaded (and once on the menu of the game Demigod).

    Recently the PSU (a Zalman ZM1000-HP) starting making a buzzing noise interspersed with periods of quiet "Buzzzz.......Buzzzzz....."

    My question is how do I tell if this is the GPU (an ATI Radeon 4870X2) or the PSU which is at fault, assuming I've ruled out a driver issue (which I'm confident I have)? I have neither replacement GPU or PSU to them with for testing.
     
  2. Silent_Buddha

    Legend

    Joined:
    Mar 13, 2007
    Messages:
    16,137
    Likes Received:
    5,074
    Substitute hardware.

    Could be motherboard. Could be CPU overheating. Could be GPU overheating. Could be PSU not providing enough power or constant "clean" power. Could be memory errors (yes failing memory can cause display driver errors).

    If you have spare parts or a friend that has spare parts or another computer you could grab spare parts from, you can try isolating the cause.

    That could prove difficult if more than one item is failing. For example I recently had a friend who had both his MB (electrolytic capacitors) and video card (overheating) failing on him. That was a royal PITA troubleshooting. And then after all that. We found 1 out of his 4 sticks of memory had gone bad.

    Regards,
    SB
     
  3. digitalwanderer

    digitalwanderer Dangerously Mirthful
    Legend

    Joined:
    Feb 19, 2002
    Messages:
    17,264
    Likes Received:
    1,781
    Location:
    Winfield, IN USA
    My first question would be was their dust in your PC when you opened it up? That sounds a bit like how my PC gets when she's due for a cleaning....
     
  4. Silent_Buddha

    Legend

    Joined:
    Mar 13, 2007
    Messages:
    16,137
    Likes Received:
    5,074
    Yeah dust could cause all of the above to overheat. MB (chipset and caps), CPU, GPU, Memory, PSU.

    Oh and forgot to mention that sound from your PSU could be the fan going out if it's more of a mechanical buzz than an electrical sounding buzz. If it's an electrical sounding buzz, that could be more serious.

    Regards,
    SB
     
  5. Quitch

    Veteran

    Joined:
    Jun 11, 2003
    Messages:
    1,521
    Likes Received:
    4
    Location:
    UK
    There was a small layer, I performed a cleanout but the problem persists.

    I'm looking to RMA the PSU as the buzzing is annoying, even if it's not a problem. Zalman have suggested an RMA, I just need to check warranty.

    Memory and CPU should be easy enough to test with memtest and Prime95 respectively, but I certainly don't have a spare motherboard, PSU or GPU :/

    GPU is idling at 78 degrees, which IIRC is actually about 8 degrees above where it normally idles. IIRC.
     
  6. digitalwanderer

    digitalwanderer Dangerously Mirthful
    Legend

    Joined:
    Feb 19, 2002
    Messages:
    17,264
    Likes Received:
    1,781
    Location:
    Winfield, IN USA
    78 is too damn high for that to be idling at, I bet you blew out your 4870x2...didn't you?

    Never, ever do that again. It's wrong. You just force the dust down into the heatsinks and they get clogged. Vacuum and toothbrush is the only way to go.
     
  7. Quitch

    Veteran

    Joined:
    Jun 11, 2003
    Messages:
    1,521
    Likes Received:
    4
    Location:
    UK
    Heh, no, I didn't start cleaning anything until after I had problems. I recall it used to idle about 70, which from what I remember is normal idle temperature for this particular GPU.
     
  8. fbomber

    Newcomer

    Joined:
    Jun 9, 2004
    Messages:
    156
    Likes Received:
    17
    I had a 4870x2 replaced in RMA, with the same problem of yours. It started like this, then it couldn´t start 3d apps without artifacts. Then, it won´t boot.
    I made tests in rivatuner with my new 4870x2. I started 3dmark vantage with temp readings from rivatuner, and, with autofan for the videocard, even if the rv770 chips don´t go over 85-86º C, other parts of the card go well above (specially the memory regulation, at 100ºC). So now, I leave, at idle, the fan in manual mode at 38%, which gives me a temp of 53-55ºC, and at load I leave, just in case, at 75%, and the maximum the memory regulation goes is 85-86ºC (other parts are cooler).
    I know the noise is annoying, but I prefer this to lose my card again, and be pc-less for another month. I bought head phones, so now the sound doesn´t bother me to much.
     
  9. Quitch

    Veteran

    Joined:
    Jun 11, 2003
    Messages:
    1,521
    Likes Received:
    4
    Location:
    UK
    I've run FurMark and the temperature didn't go above 87 degrees, holding at 83 for the most part, but shortly after I quit the app I left the room and came back to find the machine had crashed and rebooted.

    I've sent in a ticket to AMD to see what they suggest, but at this point I think I'm looking to RMA the GPU unless Prime95 turns up something with the CPU.
     
  10. fbomber

    Newcomer

    Joined:
    Jun 9, 2004
    Messages:
    156
    Likes Received:
    17
    It´s what I´m saying to you: the catalyst control center shows the temp of the chip. Not the other parts of the card. Download rivatuner, and there´s a tool in it that shows and logs the temps of several other parts of the card. Open the tool, run a demanding 3d app. Quit it, and look at your temps. You´ll be scared to use the autofan again. Just in case, put the fan at 90%, run your game, and see if the problem continues. To see if it´s overheating
     
Loading...

Share This Page

  • About Us

    Beyond3D has been around for over a decade and prides itself on being the best place on the web for in-depth, technically-driven discussion and analysis of 3D graphics hardware. If you love pixels and transistors, you've come to the right place!

    Beyond3D is proudly published by GPU Tools Ltd.
Loading...