8

I have a Windows 8.1 PC with ATI graphics (78xx series) card with up to date drivers. It's driving an old Iiyama 22" 1920x1080 50Hz LCD panel over a new DVI cable. I have two HDD, a little SSD for the OS and a couple of big magnetic discs for storage.

When playing games, my I sometimes see artefacts before the game crashes. Sometimes Windows crashes too. It happens on DirectX (up to date and under 9, 10 and 11) and OpenGL (both pre-v2 version and later).

After ensuring vents were clear, heat was under control and cleaned contacts etc, I returned the graphics card to the people I bought it from (Scan), who ran it under load (3DMark benchmark) for 24 hours without a fault being found.

Drivers are up to date and the windows dump crash report (when it takes out Windows) says that an unknown hardware fault occurs.

The rest of the machine runs fine when under load - I run artificial intelligence algorithms (no GUI) on it that run the 3 of the 4 processors at 80% and 10 of the 12GB of RAM used for 2-3 days without a problem. During that time, there's also high disk access against the magnetic discs.

If the monitor goes into standby, it sometimes omits a high pitched buzz sound after a while. I've replaced cables and isolated power.

Is it possible that a fault in the monitor is causing the graphics card to crash?

The only test I can't easily do is borrow another monitor, because I would need to borrow it for a period of time.

Edit

A co-worker has suggested I try a VGA cable, which is not plug-and-play. It's a good idea, I'll try that tonight (if the GfX card has a VGA out) and report back. VGA cable made no difference.

Tried 3DMark (it was version 11) to reproduce the test that Scan computing did, and found that it was still crashing.

Tried the onboard adapter - not powerful enough to run the same games but showed that there wasn't a problem in OpenGL.

You might also want to make sure it's actually outputting 50 Hz, and not 60, though, again, that shouldn't cause this particular kind of crash

Tried that, definitely outputing 60Hz.

10
  • Its very unlikely. A problem like this is normally caused by drivers.
    – Ramhound
    Feb 24, 2014 at 13:37
  • > ran it under load for 24 hours What kind of load? There are many different loads, some of which would not stress the same components as what you are doing with it (e.g. artefacts are often memory/vram-related, not the GPU itself - and games can load a lot of textures into memory). Are you only seeing the effects while gaming? Is it a specific game? I mean, it's theoretically possible for a monitor to cause issues, but I wouldn't expect video artefacts.
    – Bob
    Feb 24, 2014 at 13:37
  • Scan Computing ran it against 3DMark. I only see the effects when gaming and in a number of games. Does the monitor communicate its refresh rate/native res/etc? I think it must because the plug-n-play standard does. If the monitor is faulty, it could be messing with that while I play. Feb 24, 2014 at 13:44
  • 3
    @RobLang Your use of the term "plug and play" in this context is utterly incorrect and has no real meaning. Also, both DVI and VGA are able to transmit EDID information (monitor supported rates, resolutions, etc.) anyway. There have been stories of odd behaviour due to invalid EDID, but certainly not causing video artefacts only under load.
    – Bob
    Feb 24, 2014 at 13:51
  • 1
    @RobLang Have you tried running 3DMark yourself? Preferably the same version they were running. Also, since you don't have another monitor to test, consider testing a different video adapter - e.g. the onboard one. This may not be feasible if the issues only occur during gaming, though. You might also want to make sure it's actually outputting 50 Hz, and not 60, though, again, that shouldn't cause this particular kind of crash.
    – Bob
    Feb 24, 2014 at 16:52

2 Answers 2

4

The solution was:

After performing a CC Clean of the drivers (back to Windows default, all folders removed etc), I used the Catalyst Control Center driver update tool to check for the best driver and download it. Problem persisted. CC Clean again.

For fun, I then used Windows Driver Update (Control Panel->Devices->Display Driver->Properties->Update Driver) and that downloaded a different version of the driver - which is actually the latest.

Cause

The ATI Control Centre driver check update tool was not downloading the latest driver for my card, it was downloading the previous one. Have raised a bug report with the ATI CC team.

Thank you everyone for your help, I am somewhat flabbergasted that the Windows Driver Update would get a newer version than the ATI CC tool.

1

I suppose it is possible a monitor could cause a crash with a driver, but its very unlikely. Most graphics problems are driver and graphics card hardware issues.

Obviously, the first thing to try would be a different monitor. If you still have issues with a different monitor, then its the card or driver. If the problem cleared up (which I highly doubt), then it would be the monitor.

You said you are using the most up to date driver. Newer isnt always better. Its not uncommon for newer drivers to introduce bugs. Try downloading and installing an older driver, preferably one that has WHQL certification.

Check the cards temperature. Perhaps its overheating. If possible test the graphics card in another machine. Perhaps it isnt getting enough power or airflow.

PS:

I also want to add that you said you see artifacts. Artifacts are commonly due to a card overheating and/or failing. Since the card was supposedly tested, then it sounds like overheating.

2
  • On installing drivers - make sure the old one is fully removed. Sometimes this might require multiple uninstall + reboot loops until it's completely off the installed program list (sometimes, old ones reappear after a reboot...)
    – Bob
    Feb 24, 2014 at 17:20
  • I used the CC cleaner to completely remove the drivers back to 640x480. Temperatures all good. Feb 24, 2014 at 22:02

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .