Random BSOD. Grahpics driver?

[Surgikill]

Hey guys. Have been having some BSOD's. I don't know if it is graphics driver related or related to something else. Here's the SF diagnostic tool file and I also did a debug and came up with this text file as well.

 

[Golden]

************* Symbol Path validation summary **************
Response                         Time (ms)     Location
Deferred                                       SRV*C:\SymCache*http://msdl.microsoft.com/download/symbols
Symbol search path is: SRV*C:\SymCache*http://msdl.microsoft.com/download/symbols
Executable search path is: 
Unable to load image Unknown_Module_80000000`08deee6a, Win32 error 0n2
*** WARNING: Unable to verify timestamp for Unknown_Module_80000000`08deee6a
*** ERROR: Module load completed but symbols could not be loaded for Unknown_Module_80000000`08deee6a
Debugger can not determine kernel base address
Windows 7 Kernel Version 7601 (Service Pack 1) MP (4 procs) Free x64
Product: WinNt, suite: TerminalServer SingleUserTS Personal
Built by: 7601.17592.amd64fre.win7sp1_gdr.110408-1631
Machine Name:
Kernel base = 0xfffff800`0324a000 PsLoadedModuleList = 0xfffff800`0348f650
Debug session time: Wed Dec 24 10:25:07.555 2014 (UTC + 10:30)
System Uptime: 0 days 0:02:47.602
Unable to load image Unknown_Module_80000000`08deee6a, Win32 error 0n2
*** WARNING: Unable to verify timestamp for Unknown_Module_80000000`08deee6a
*** ERROR: Module load completed but symbols could not be loaded for Unknown_Module_80000000`08deee6a
Debugger can not determine kernel base address
Loading Kernel Symbols
.
Loading User Symbols
*******************************************************************************
*                                                                             *
*                        Bugcheck Analysis                                    *
*                                                                             *
*******************************************************************************

Use !analyze -v to get detailed debugging information.

BugCheck A, {0, 2, 1, fffff800032ced92}

***** Debugger could not find nt in module list, module list might be corrupt, error 0x80070057.

Probably caused by : Unknown_Image ( ANALYSIS_INCONCLUSIVE )

Followup: MachineOwner
---------

I'm afraid that .dmp file is no use at the moment.

Please do the following:

Run Driver Verifier for 24 hours or the occurrence of the next crash, whichever is earlier.
http://www.sevenforums.com/tutorials/101379-driver-verifier-enable-disable.html

Driver Verifier will cause your computer to run very sluggishly - this is normal. What it is trying to do is force your system to BSOD and isolate the offending driver/s. When it does, reboot, disable driver verifier, reboot as normal and upload the new dmp file/s here.

I recommend creating a system restore point before turning on driver verifier:
http://www.sevenforums.com/tutorials/697-system-restore-point-create.html

If your system fails to boot to desktop once driver verifier is enabled, turn it off by booting into Safe Mode:
http://www.sevenforums.com/tutorials/69585-safe-mode.html

 

[Surgikill]

It crashed again. Here's the file.

 

[Surgikill]

Any thoughts? I'm pretty sure it's related to my sli.

 

[Golden]

*******************************************************************************
*                                                                             *
*                        Bugcheck Analysis                                    *
*                                                                             *
*******************************************************************************

Use !analyze -v to get detailed debugging information.

BugCheck 116, {fffffa800c163010, fffff88005bf4bfc, 0, d}

*** WARNING: Unable to verify timestamp for nvlddmkm.sys
*** ERROR: Module load completed but symbols could not be loaded for nvlddmkm.sys
Probably caused by : nvlddmkm.sys ( nvlddmkm+13ebfc )

Followup: MachineOwner
---------

It appears to be the case - 0x116 bugchecks occur when Windows has trouble switching from one display to another within an allocated time frame.

 

[Surgikill]

So what would that be?

 

[Golden]

Unfortunately, I have no idea how to overcome that. I'll see if I can find some help for you.

 

[essenbe]

Graphics cards are pretty simple in a desktop rig. It is either the driver, the PSU the motherboard or the card. Try Reinstalling the most current WHQL driver using this method http://www.sevenforums.com/tutorials/327414-nvidia-drivers-avoid-problems.html

If that does not resolve the problems try stress testing with only one card installed at a time. http://www.sevenforums.com/tutorials/100356-video-card-stress-test-furmark.html

What temp monitoring program are you using and what temps are you seeing? Before you run Furmark, make sure you have a very good temp monitoring program and stop the test if they get too high.

 

[Surgikill]

I'm also having a problem where my keyboard spontaneously disconnects. It's not the cable or the keyboard, I tested both on a separate system. Also, while this happens all my other usb devices lag. I'm going to reinstall windows and see if I can't fix this.

 

[essenbe]

Surgikill, You never told us the results of your latest test. Please let us know and try not to get ahead of us. If you have a bad graphics card, reinstalling Windows will not help.

If you are overclocking anything, please set it back to default levels.

 

[Surgikill]

@essenbe When I go to stress test SLI the two cards have different stock clocks. Should I underclock the higher card, overclock the lower card, or leave them the same? Also how long should I run furmark?

UPDATE: So I ran furmark on my evga card which is a gtx 460 made by EVGA. When I installed my 2nd card, which is a gtx 460 made by Galaxy and is super OC the computer booted and it reverted back top 800x600 resolution and searched windows update. Also, before the post screen I noticed a black screen that said something along the lines of NVIDIA Corporation GTX 460. I never saw this with my other card.

UPDATE 2: So I just ran furmark on my 2nd card. It leveled out at 82C, about 8C hotter than the first card. No issues, both run at same settings. Now I'm going to install both and run furmark in SLI.

UPDATE 3: So I installed both cards and when I booted up my machine it told me my sli bridge was missing. Clearly isn't (check pics) SO I am assuming that my sli bridge is shot? I'm going to try cleaning it and see if that helps.

 

[essenbe]

Everything sounds normal except the SLI Bridge. Do you think your Bridge could be bad?

 

[Surgikill]

Everything sounds normal except the SLI Bridge. Do you think your Bridge could be bad?

I got the bridge working. I remembered that if I pushed the bridge completely all the way in on each card it would not work. I ran furmark on both cards at 720mhz and they were stable. Just tried it at 810 and I got a bsod. I think the top card went over 100c. The bottom card was around 80c. I'm going to try furmark again and watch it closely.

EDIT: I also am not noticing an fps increase on furmark with sli. Is this normal

EDIT 2: just crashed again. Was at 99C. Even at 720mhz it gets to this temp. Should I remove the plastic shroud?

 

[essenbe]

Doesn't the plastic shroud hold the fans on? What you need would be something like Nvidia Precision X or MSI Afterburner and set a custom fan profile and add some more fans blowing on the Graphic cards. Try taking the side off the case and blowing a house fan directly into the case on the cards and see if that helps. Those cards don't need to be anywhere close to 100C.

 

[Surgikill]

Doesn't the plastic shroud hold the fans on? What you need would be something like Nvidia Precision X or MSI Afterburner and set a custom fan profile and add some more fans blowing on the Graphic cards. Try taking the side off the case and blowing a house fan directly into the case on the cards and see if that helps. Those cards don't need to be anywhere close to 100C.

Well I use Asus GPU tweak and set both fans to clock at 100% @>90C

Anywho I decided to modify the exhaust port on the evga card. I got rid of all the slots so there would be more airflow. In the process I decided to re-apply thermal paste. When I took the cooler off and looked at the die I couldn't believe what I saw. There was solder on the die, and flux all around the edges. I figured "Oh well, shit happens" So I took and applied a very generous glob of arctic silver and went to put the cooler back on. WELL when i put it back on and I looked at it through the side of the card I could see light between the cooler and the die, and I put enough thermal paste on to at least cover 2 cpu dies. I could literally fit 4 index cards between the cooler and the die. So I modified the studs on the cooler so it would actually touch the die. Now I'm running furmark again and I'm seeing slightly better temps. 95C on stock vs 98C, but it's still quite warm. I will put my side panel on with the fan and see if that helps.

I'll also post some pics in a minute.

 

[Surgikill]

Here's the link to the pictures. They're too big to upload directly.


https://www.dropbox.com/sh/uxkurclz76wklv7/AAAo6t7_bkV4_1x3czvYAgw5a?dl=0

 

[essenbe]

I see, but I think I would still leave the shroud on.

 

[Surgikill]

I see, but I think I would still leave the shroud on.

The shroud is on. I think I'm not getting enough airflow through the card because it is in SLI. Do you have any recommendations for lowering temps?

 

[essenbe]

That's a common problem with SLI/CFX,. The only solution I know of are more, well placed fans or water.