*FIXED* I just lowered my GPU speed back down to a generic 260 stock and it's fine. Thanks to Instagib! Hey guys, Whenever my PC is under any medium-high graphical load it crashes and needs to be hard restarted. This occurs when I try to play Team Fortress 2 and Starcraft 2 and also when I run Furmark at a high resolution and full screen. The general theme is, the higher the graphics settings and detail, the shorter it lasts. For e.g. If I run Furmark @ 800x600 then it can cope for a long while, but when I up the resolution to 1440x900 and use full-screen, it crashes within 5 seconds. Strangely, I can play less intense games like Counter-Strike 1.6 and Quake Live for as long as I want and they never crash! I don't think its my actual graphics card as I have tested it in another PC and it ran Furmark @ 1600 resolution and full-screen for over 2 minutes (before I stopped it.) Just as it crashes, my CPU temperature is around 60c and GPU goes up to about 70c so it can't be overheating either. Things I have tried: I tried 4 different graphics card drivers I have also tested each ram stick individually and in each motherboard slot; all to no avail. I eventually reformatted but the problem still persists. Specs: Intel Core 2 Duo 6600 ASUS P5N-D 4GB(2 pairs) Corsair DDR2 6400 XFX NVidia GeForce 260GTX Black Edition Xclio X14 600W Onboard sound Windows XP SP3 I hope you guys have some ideas and if you need any more information, just let me know Thanks in advance Edit: Example pic of the crash
Can you pinch the PSU from the other PC and see if that fixes the problem? In light games your GTX 260 probably isn't running flat out so it draws less power. When it is running at 100% it may be requiring more than your PSU can handle. Judging by the few reviews I can find the PSU you have is pretty decent but that doesn't rule it out as the source of the problem.
Furmark uses a lot more power than most games would hence why it is used for stress testing. Correct me if i'm wrong but judging from the specs for the only 600w PSU they advertise on the Xclio website, each of your 4 rails is limited to 18A The manual for that PSU doesn't seem to give any information as to how the rails are distributed. How is it wired up to your graphics card? Are you using 2 PCIe connectors on the same wire? If this is the case then try connecting another PCIe connecter so that they are from different wires and hopefully from different rails.
Hi guys, Apologies for taking so long to get back to you all - I've only just had time to pickup a new PSU (Antec True Power 650W) and unfortunately... it's still crashing! Please please please if you have any other advise I will be so grateful. I'm running out of ideas! Thanks in advance
Another update (hope this helps): 1. I have uninstalled my previous gfx drivers (safemode, uninstaller, driver sweep) and reinstalled the very latest ones. Still crashing 2. I have swapped the card to the 2nd PCI-E slot and still crashing. 3. I have also checked all the caps on the mobo and they are fine. I have had some success though. I left the PC hanged on a nice purplely green corrupt screen for a few minutes, to see what would happen. The sound would loop and the game would almost run at 1 frame per 20 seconds whilst the gfx are totally whacked. BUT THEN! the game crashed to desktop! I could see the error messages and i print screened them. Edit: The other error was generated from half-life 2 engine Note: This isn't EXACTLY what I saw. On the screen was strange black/white train track lines running vertically in stripes which disappeared on reboot, but didn't appear on screenshot.
Even though i tried the card in a friend's computer and it ran Furmark @ 1600 for 2 minutes without failing? Image of the lines on my screen (but not on the screenshot)
Can it run 100 loops of the Crysis benchmark? That loks like what my 8800GT's used to do before they died. At that point it's usually RMA time.
Before you go out and get a new card, download MSI Afterburner and drop the gpu speeds back to stock. It might be possible that the factory overclock has become unstable. It's a longshot though; it does seem like the card is on it's way out.
What is the GPU reported temps in the friends machine when it runs consistently over a minute? Just because the GPU temp is ok, though, doesn't necessarily mean all is well but inbstead it's an indicator that maybe not all the undesirable heat is being extracted ok. Location of the PC? Air outflow being blocked maybe? Even air being drawn in from near a heated source? What's the difference (if any) in the ambient temp in the cases? Taking a lead from Saspro's suggestion, it may simply be the case that the gpu ram sinks are not conducting correctly and therefore the extra couple of degrees C in one case is making a difference over another. For running in the friends machine, I'd suggest stressing it for longer than 2 minutes. If he'll let you, try running a continuous test for an hour or so on it.