GTX 660 Graphic Driver Issues - Looking for suggestions

[Malzian]

I have a GTX660Ti, was getting some pretty intense tearing for a while until I capped my frame rate at 60fps. Occasionally I'll get a weird flash of color, but mostly I think that's been due to scenery. Also try turning on ambient occlusion if you haven't, really helped my performance out.

[Dainius56]

almost definitely sounds like an unstable overclock, overheating, or needing to do a clean install of drivers. Clean install is very important when going from driver to driver looking for better performance

[ZeroEx]

Yeah, sounds like overheating GPU.
Had some serious glitches on my Laptop, and it was because GPU was overheating.

[Segraine]

My gpu is at stock and it only gets up to 55 C when under a load (using Nvidia's System Monitor). I always clean install drivers. Learned that a long time ago! Tried installing 326.19 but still experienced crashes in dungeons. This time, I finally saw a blue screen! It seems nvlddmkm.sys is throwing the error (DRIVER_IRQL_NOT_LESS_OR_EQUAL). Locking the framerate to 60 fps has helped my stability but dungeons are still a problem. Again, the open world rarely crashes, which is odd considering there is more to draw.

Here are screens of what is happening:




I thought it was a heat problem too, but no other game causes this problem.

[Geostigma]

I have the same issue playing only in windowed mode, no matter what graphic resolution. My fix was to play in fullscreen. It's a little more tedius, but it works (just a hard flash when alt tabbing). I had no issues in windowed mode with my gtx570. But with my 660 and my 760, both have tearing in windowed mode. Again, fullscreen fixed it without any issue, and actually I have smoother playback.


This was of course, during beta and early access only. As I have not been able to play since Aug 24 because of service account/customer support congestion

edit: after reviewing your screenshots, my tearing was not this significant, not by any means. 55c IS hot, but falls in the category "slightly above average" i would try to, if your expertise allows, remove the heatsink, apply some new thermal compound (preferably silver), and remount the heatsink/fan (you may even consider a new fan). I would try to push it to 45c. If that amount of tearing still persists, you are most likely looking at a developed problem due to gpu damage from sustained heat.

[Raist]

hmm.... wonder what kind of state the power supply is in too. Is it an 80+ certified model (may not have 80+ logo, but maybe a Xfire or SLI certified logo)?

Another thing you might want to look for is the power profile for the card. Forge texactly what and where it was located in the nVidia settings, but there is a section somewhere in there about Adaptive or Performance mode for it's power consumption. If it's on Adaptive, it might not be kicking in/out of high power/performance state properly.

[JamieLee]

irq not less then equal usually means ram related issue or theirs a problem with the north-bridge controller on your motherboard.

IRQL_NOT_LESS_OR_EQUAL error in Windows 8 is displayed in Blue Screen of Death (BSOD). Just like other Blue Screen of Death errors, IRQL_NOT_LESS_OR_EQUAL error also arises due to corrupt software application files, corrupt system files, incompatible or inappropriate device drivers, corrupt device drivers, faulty hardware devices, etc.

Typically IRQs are the interrupt requests that are sent to the microprocessor via south and north bridges. If IRQL_NOT_LESS_OR_EQUAL error occurs, chances are that an issue with the application or driver is related to the memory or interrupt request.

For Driver irql not less then equal it could be hardware overheat... It could be the card somehow got dislodged from the slot (reseating cards in the case sometimes fixes issues) I have personally had a video card some how come up out of the pci e slot just enough to throw windows into a shit fit.

It could be corrupt system files.. Or a in comparable program or a conflict between a virus scanner and the program itself..

you can also try a manged start-up and disable all the apps on start-up and try it if that works re enable apps tell you find one that causes it..

clean install of windows and reload can also help.

[Wren_Fuoren]

314.22's are stable too

[maTyaR]

Crazy artifacts! You may try using the videocard on another computer (if you have one). If you don't, try playing another game and see if it causes the same issue to make sure it's just not the game itself.

A more possible error could be your videocard VRAM itself that is defective.

Hope you find a solution!

[stuntdubs]

55 is actually really cool under full load for a 660. Anything under 70 is considered fine, 70 - 80 is pushing it, and 80+ is too hot.
I play at1080p high quality on my 660 and it hovers around 65-70. never have any problems though.
It certainly looks like a classic heat problem. As said above could be faulty ram.

What are the rest of your system specs? (specifically motherboard, cpu and powersupply)
I see you mentioned playing LFD2 and shogun 2, are you able to try somemore graphically intensive games, max them out and see if you get the same problems. Something like one of the crysis games/battlefield 3 etc would do the trick.

Reseat your gfx card intoyour motherboard.
Try dial back your core and memory clocks on the card slightly.
Drivers you've obviously tried.
If possible run the game on a different system with your gfx swapped into it (alot of trouble I know)
Don't run in windowed.
Manually set your GPU fan to max before playing (75 I think)
Check your CPU temps
Disable all system monitoring software (precision x, afterburner etc) if you have any.

Good luck