Announcement

Collapse
No announcement yet.

Getting video memory corruption

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Getting video memory corruption

    Hi all,

    Have run Burnintest, after I found out that my graphics card might be a trouble.

    Symptoms:
    - When I play Modern Warfare 2, when a multiplayer match starts, the screen colors go nuts. Last time I had half of the screen filled with a pink color, and then I got a BSOD indicating this error: STOP 00x00000116 nvlddmkm.sys (which belongs to nvidia).
    Sometimes I get blue strokes/lines all across the screen, then it freezes and crashes, also after having tried to play MW2.
    - Recently when I restarted, the bios text is colored differently than usual, with some light yellow colors. But not on every reboot.

    So I ran Burnintest, on all categories, and after 15 minutes it shoed over 3000 errors in the 2D test appointing it to video memory corruption. Had no errors in any of the other test-categories.

    Below the testlog is shown:

    Recently I changed my powersupply to a new one, since the old one started to make crackling noises. And I actually ran the computer with this for some hours. Don't hope the powersupply made some damage to the rest of the pc, including the graphics card.

    Please help me identify the problem


    **************
    RESULT SUMMARY
    **************
    Test Start time: Wed Jan 20 16:44:21 2010
    Test Stop time: Wed Jan 20 16:59:23 2010
    Test Duration: 000h 15m 02s

    Test Name Cycles Operations Result Errors Last Error
    CPU 146 462 Billion PASS 0 No errors
    Memory (RAM) 1 8.088 Billion PASS 0 No errors
    2D Graphics 8 8256 FAIL 3456 Video memory corruption
    3D Graphics 51 02231 PASS 0 No errors
    Video Playback 29 872 PASS 0 No errors
    TEST RUN FAILED

  • #2
    There are some past posts on this issue.
    http://www.passmark.com/forum/showthread.php?t=244
    http://www.passmark.com/forum/showthread.php?t=262
    http://www.passmark.com/forum/showthread.php?t=2331

    In summary you can ignore a few of these errors. If you get a lot of them, then you should be more concerned.

    In this case 3000 errors in 15min is a lot. Especially combined with the other visual artifacts you are seeing.

    The most likely scenario is that the video card RAM is bad. I would initially try swapping the video card.

    Comment


    • #3
      I have run Furmark to test the GPU and only in 'stability mode' does the graphics card crash the pc, but only when reaching high temperatures above 75 degree.

      Also I ran Speedfan to test temperatures and voltages:
      Temperature under windows is 56 degree. seems fine.
      Voltages:
      Vcore1: 0,86V
      Vcore2: 1,58V
      +3,3V: 3,36V
      +5V: 4,95V
      +12V: 3,58V (just seconds ago it was 1,60V, then 1,47V, then 1,15V)
      -12V: -1,01V
      -5V -8,73V
      +5V: 3,66V
      Vbat: 3,22V

      I don't like the 12V readings.. What does that mean?

      I will try and swap the graphics card tomorrow and run BIT again.

      Comment


      • #4
        About the voltage results: I recently switched to a new powersupply, because the old one broke. It made crackling sounds..

        So I would not understand if the new powersupply is making somekind of errors.

        Btw I run:
        Gigabyte P55A-UD3
        Core i5 750
        4gb ram
        Geforce 8800GTX

        Comment


        • #5
          I have now double checked the +12V voltage in CPUID HWMonitor, and it also reads about +1,15V..

          That's quite a lot away from +12V.

          Edit: Just rebooted into bios to see bios voltage results. The +12V is at 12,35V.

          So bios is hopefully telling the right thing?
          Last edited by piopakk; Jan-20-2010, 10:51 PM.

          Comment


          • #6
            I just ran BIT again. This time only to check the graphics card. Did the 2D, 3D and video test. After 15 min I got no errors at all.

            Isn't that strange?

            GPU temperature after this test was 57 degree and never ran higher within those 15 minutes.
            Can't remember the GPU temperature when I ran the first test with the 3000 errors.

            Comment


            • #7
              Run a third BIT test immediately after the second run. Same test categories with 2D, 3D and video, though added CPU and RAM. After the 15 min. still no errors.

              How come I can get over 3000 errors and then none?

              ps. Sorry for all my replies, just thought it would be nice with added info's to my problem..

              Comment


              • #8
                Yes, i think those other tools are reporting incorrect voltage values. If what they reported was true, then I don't think the machine would even boot.

                As you point out, it might be temp related.

                Comment

                Working...
                X