Hello,
I recently received a 2nd 680 GTX to turn my setup into SLI. (Also purchased a coolermaster 750w psu to upgrade from my old 550w psu, to support the new power demand).
I am having some issues with the cards in SLI that I have tried some hard google searching on but solutions seem varied and few and far between.
The main problem is that during fullscreen applications and even just sitting on the desktop and even just on the desktop not doing anything (standing up away from the pc) the screen will go black as if it went out, then turn back on as if it lost connectivity. Very briefly a display error message will flash in the taskbar saying "Display driver has stopped responding and has recovered
Display Driver NVIDIA Windows Kernel Mode Driver, Version 314.22 stopped responding and has successfully recovered.". And sometimes I will also receive an error message in the middle of the screen saying that there was an OpenGL crash with "Error Code: 3".
This error can happen at any time with no obvious trigger. I found that I can force the problem (however this may be something different / just general overheating problem) by letting the graphics card get above 85c by turning off any fan controls and letting it try and run on its own stock fan speed. However with Evga Precision X on and my own custom fan curve I never hit those temps yet I still get crashed from everywhere between 30c and 60c idle/load temps.
When this error happens in fullscreen games the monitor goes out and then the display returns but the game will be running at a fraction of the performance and be extremely slow and choppy. For example, playing CS:GO (Counterstrike Global Offensive) I have 300-400 fps standard, then the error will hit, monitor comes back and game returns but fps has dropped to 90 or lower but it feels like its at 15fps, extremely laggy and totally unbearable. It does not recover from this state.
This does not happen when it crashes/errors from the desktop, everything returns to normal.
A second and minor issue/problem that confuses me is that I can only get the computer to detect 2 graphics cards when they are arranged in a certain order in the PCIe slots. Swapping them around results in only the card in the 1st PCIe slot to be recognised. I have tried this around a few ways and have tried running each card individually in the first slot inbetween swaps and this issue persists.
What I have tried in regards to troubleshooting:
- Heavy temperature monitoring to see if temps were peaking causing this. Discovered this is not the case as mentioned before as it can happen at any temperature.
- Double checking power output on voltage rails, there are no problems in this area, all the rails are very balanced/accurate and their is no heavy power draw under a stress test on the cards with evga oc scanner test.
- Swapping the cards around (gives previously mentioned problem)
-- Odd thing is that the only configuration it works in is the new card in the PCIe 1 slot and the old card in the PCIe 2 slot. Having new card in slot 2 and old card in slot 1 shows only old card in 1 and nothing in 2, no hardware detected.
- Tried each card individually under 3dmark11 test and individual tests on Battlefield 3 on all Ultra settings. To see if it was either card causing the issue.
-- Noticed that the NEW card did error 2x across a long period but it was much harder to reproduce than when the 2 cards are in SLI. But the OLD card performed perfectly, as expected as I never had this issue before bringing a 2nd card into the mix.
- Tried fully uninstalling drivers from each card individually and reinstalling the 314.22 but this didn't change anything.
Other specs are:
p8p67 pro rev 3.1 mobo
i5-2500k @ 4.3ghz
Corsair Dominator (2x4GB) DDR3 PC3-17100C9 2133MHz
750w psu
both 680gtx are the 2gb version
My current thoughts are that it could be the new card that has fault(s) and that I should return it to the seller and get a new one and go from there. This is because it did produce the error when the card was running on its own although much harder to make it do so.
I'd like to know if there's any knowledge on this sort of error with SLI setups. How to troubleshoot a "windows kernel driver mode error" or OpenGL error.
Why will the cards only work in one certain PCIe configuration?
What do I do, if anything, to stop this error/crash because it makes all games/fullscreen apps unusable / unbearable if they could just crash and performance drop any minute.
I recently received a 2nd 680 GTX to turn my setup into SLI. (Also purchased a coolermaster 750w psu to upgrade from my old 550w psu, to support the new power demand).
I am having some issues with the cards in SLI that I have tried some hard google searching on but solutions seem varied and few and far between.
The main problem is that during fullscreen applications and even just sitting on the desktop and even just on the desktop not doing anything (standing up away from the pc) the screen will go black as if it went out, then turn back on as if it lost connectivity. Very briefly a display error message will flash in the taskbar saying "Display driver has stopped responding and has recovered
Display Driver NVIDIA Windows Kernel Mode Driver, Version 314.22 stopped responding and has successfully recovered.". And sometimes I will also receive an error message in the middle of the screen saying that there was an OpenGL crash with "Error Code: 3".
This error can happen at any time with no obvious trigger. I found that I can force the problem (however this may be something different / just general overheating problem) by letting the graphics card get above 85c by turning off any fan controls and letting it try and run on its own stock fan speed. However with Evga Precision X on and my own custom fan curve I never hit those temps yet I still get crashed from everywhere between 30c and 60c idle/load temps.
When this error happens in fullscreen games the monitor goes out and then the display returns but the game will be running at a fraction of the performance and be extremely slow and choppy. For example, playing CS:GO (Counterstrike Global Offensive) I have 300-400 fps standard, then the error will hit, monitor comes back and game returns but fps has dropped to 90 or lower but it feels like its at 15fps, extremely laggy and totally unbearable. It does not recover from this state.
This does not happen when it crashes/errors from the desktop, everything returns to normal.
A second and minor issue/problem that confuses me is that I can only get the computer to detect 2 graphics cards when they are arranged in a certain order in the PCIe slots. Swapping them around results in only the card in the 1st PCIe slot to be recognised. I have tried this around a few ways and have tried running each card individually in the first slot inbetween swaps and this issue persists.
What I have tried in regards to troubleshooting:
- Heavy temperature monitoring to see if temps were peaking causing this. Discovered this is not the case as mentioned before as it can happen at any temperature.
- Double checking power output on voltage rails, there are no problems in this area, all the rails are very balanced/accurate and their is no heavy power draw under a stress test on the cards with evga oc scanner test.
- Swapping the cards around (gives previously mentioned problem)
-- Odd thing is that the only configuration it works in is the new card in the PCIe 1 slot and the old card in the PCIe 2 slot. Having new card in slot 2 and old card in slot 1 shows only old card in 1 and nothing in 2, no hardware detected.
- Tried each card individually under 3dmark11 test and individual tests on Battlefield 3 on all Ultra settings. To see if it was either card causing the issue.
-- Noticed that the NEW card did error 2x across a long period but it was much harder to reproduce than when the 2 cards are in SLI. But the OLD card performed perfectly, as expected as I never had this issue before bringing a 2nd card into the mix.
- Tried fully uninstalling drivers from each card individually and reinstalling the 314.22 but this didn't change anything.
Other specs are:
p8p67 pro rev 3.1 mobo
i5-2500k @ 4.3ghz
Corsair Dominator (2x4GB) DDR3 PC3-17100C9 2133MHz
750w psu
both 680gtx are the 2gb version
My current thoughts are that it could be the new card that has fault(s) and that I should return it to the seller and get a new one and go from there. This is because it did produce the error when the card was running on its own although much harder to make it do so.
I'd like to know if there's any knowledge on this sort of error with SLI setups. How to troubleshoot a "windows kernel driver mode error" or OpenGL error.
Why will the cards only work in one certain PCIe configuration?
What do I do, if anything, to stop this error/crash because it makes all games/fullscreen apps unusable / unbearable if they could just crash and performance drop any minute.
