Solved BSOD Random but frequent occurence

[Ken Cravillion]

New to the forums. I have a custom built machine built by a friend that I've been using for the past two + years. Within the last couple months I have been getting frequent BSOD's. It's happened while internet browsing, just looking at open folders, while working in Photoshop Lightroom and even using the Canon print to CD program.

I use this computer primarily for photo editing as that is my full time business.

I can describe when a BSOD is coming. First I'll notice a bunch of small horizontal lines near the mouse pointer in a small square area. Then the display will become jagged and pixelated and shortly after that I often get the BSOD. I've thought that the video card could be going bad or something with the driver so I reinstalled that with no effect. I've also updated the BIOS also with no effect.

My system configuration is in my profile.

 

[essenbe]

Hi Ken Cravillion, I'll try to help you. I've looked at your dump files as well as the error reports. what I find in the last 3 are what you expected.

 atikmpag
start             end                 module name
fffff880`08329000 fffff880`0837e000   atikmpag T (no symbols)           
    Loaded symbol image file: atikmpag.sys
    Image path: \SystemRoot\system32\DRIVERS\atikmpag.sys
    Image name: [COLOR=red]atikmpag.sys[/COLOR]
    Timestamp:        [COLOR=red]Wed May 23 21:46:18 2012[/COLOR] (4FBDA0FA)
    CheckSum:         0005835D
    ImageSize:        00055000
    Translations:     0000.04b0 0000.04e4 0409.04b0 0409.04e4

That is your Graphics Drivers/Card and it shows a TDR error . A TDR error is a timeout detection and recovery. What happens is the video driver fails and does not recover in the time allotted by Windows so Windows determines it is a video card failure. The driver appears to be pretty old and was from your last dump file.

Another one is apparently a raid issue, at least the dump files indicate that. I don't notice Raid mentioned in your system specs. But, the driver mentioned was

mvs91xx
start             end                 module name
fffff880`04346000 fffff880`043d5000   mvs91xx  T (no symbols)           
    Loaded symbol image file: mvs91xx.sys
    Image path: \SystemRoot\system32\DRIVERS\mvs91xx.sys
    Image name:[COLOR=red] mvs91xx.sys[/COLOR]
    Timestamp:        Mon Sep 19 06:31:51 2011 (4E772827)
    CheckSum:         00050B0E
    ImageSize:        0008F000
    Translations:     0000.04b0 0000.04e4 0409.04b0 0409.04e4

That is your Marvell 91xx SATA Controller Driver. If it is a motherboard port you should be able to update it from your Motherboard's web site, Motherboards - P9X79 PRO - ASUS

The one that concerns me most is this one. If this is the case, it can make the others, at least inaccurate.

SYMBOL_NAME:  nt!MiDeleteAddressesInWorkingSet+27f
FOLLOWUP_NAME:  MachineOwner
MODULE_NAME: nt
DEBUG_FLR_IMAGE_TIMESTAMP:  531590fb
IMAGE_VERSION:  6.1.7601.18409
IMAGE_NAME:  [COLOR=red]memory_corruption[/COLOR]
FAILURE_BUCKET_ID:  X64_0x50_nt!MiDeleteAddressesInWorkingSet+27f

I realize this is very inconvienent especially for one who uses his computer to make a living. But the test for this will probably take you 20+ hours to run. It has to be run for 8 complete passes all at once. There is no other way to test it. The test I am going to suggest is the best test we have for testing ram.

Please Run Memtest86+

   Information

Please download from this site only http://www.memtest.org/ in the middle of the page are the Download links, you can download the ISO.zip or the Auto USB Flash Drive installer.zip

Extract the Zip file. If you chose the ISO image, burn it to a CD using Windows Disk Image Burner or any Image burner you may have. If you downloaded the Auto USB installer, extract it, insert your USB 2.0 Flash Drive and take note of the drive letter. Run the installer, select the Flash Drive Letter, check the format box and press next. It will install memtest86+ to a flash drive. You can use either V4.20 or V5.01. Boot from your selected media. If you use V5.01 it will tell you to press certain buttons at the start, please press no buttons. The test will begin on it's own and continue to run until you stop it. It needs to run for 8 complete passes or until you receive an error. If you receive an error, stop the test. Even 1 error is a fail. Each pass tests a different part of the ram and each of the 10 tests in each pass tests something different. It takes a minimum of 8 passes to completely test the ram, more passes are better. It is quite a long test and will take several hours depending on how much ram you have. Due to the time length it is best to run overnight. If you have any questions, please do not hesitate to ask. I am more than willing to hear your comments or questions and explain the best I can.​

 

[Ken Cravillion]

Thank you very much! Would you suggest I start with the memory and go from there?

Regarding the graphics driver the one installed is the latest one according to ATI's website. There is no newer driver for that card.

The two WD 2TB HD's are mirrored as that is where my images are backed up. That's why you see RAID. I didn't set it up my builder did.

 

[Ken Cravillion]

For the graphics driver this is the latest:

AMD FirePro? AMD FireMV? Unified Display Driver

Current as of June of 2012...

(EDIT)

Doing some more digging I found more recent drivers:

http://support.amd.com/en-us/download/workstation/previous?os=Windows 7 - 64

 

[essenbe]

I know it's a PITA doing memtest. I've been there numerous times, but ram will ruin everything. So, I would start there. Also, if you are overclocking anything, set everything back to defaults. As you can see from my sig, I am not opposed to overclocking, I'm overclocked right now. But, it can become unstable and throw some strange errors. So, if you are, please set it to defaults until we get it stable, then overclock it like you want. Then if you start crashing, you know it's the overclock. That's exactly what I do, if I can't figure it out. So, yes let's do the memory test first.

 

[Ken Cravillion]

Just to update... I'm going to be running the memtest this weekend (have the ISO on CD ready to go). I wanted to get last weekends photography job finished.

In the mean time I had two more BSODs. Here are the latest two logs...

 

[essenbe]

here's what they say

Use !analyze -v to get detailed debugging information.
BugCheck 50, {fffff6fffe7e7e78, 0, fffff800030f8401, 2}
 
Could not read faulting driver name
Probably caused by : [COLOR=red]memory_corruption[/COLOR] ( nt!MiDispatchFault+d1 )
Followup: MachineOwner

AtihdW76
start             end                 module name
fffff880`0a636000 fffff880`0a651000   AtihdW76 T (no symbols)           
    Loaded symbol image file: AtihdW76.sys
    Image path: \SystemRoot\system32\drivers\AtihdW76.sys
    Image name: AtihdW76.sys
    Timestamp:        Thu [COLOR=red]Dec 19 21:15:49 2013[/COLOR] (52B3B665)
    CheckSum:         0002221E
    ImageSize:        0001B000
    Translations:     0000.04b0 0000.04e4 0409.04b0 0409.04e4

ATI Function Driver for HD Audio Service

 

[Ken Cravillion]

Just an update. The errors got really bad a few days ago. Finally yesterday the video card failed. It wouldn't power up either monitor. So I shut down the computer and installed a new card. After installing the new card I ran the computer for 9 hours yesterday (doing usual tasks that would cause BSOD's which was everything) and have had it on for 3 hours so far this morning. Happy to report no issues as of yet and crossing my fingers that this issue is done. I'm going to give it a few days to make sure before closing this thread.

 

[essenbe]

That's great news. I hope your good luck continues. I appreciate the update. Pleas give us another in a few days. That could have very well been the problem all along. Very well done. Whenever you feel the problem is solved, please mark the thread solved. But also let us know how it goes for you. Great job.
By the way, it's quite possible that the memory corruption was the graphics card memory. That is not the usual meaning, but I have seen it be that several times.

 

[Ken Cravillion]

I gave it another 5 days and all is well. No errors to speak of and I've been running the machine pretty much every day. Thank you for taking the time to help!

 

[essenbe]

No problem, I'm just glad it is running right. Good job.