Solved Video card BSOD ATIkmdag.sys 0x0000003b

[computergeekguy]

Hey guys, my computer blue screens approximately 30-45 minutes into gameplay of Battlefield 3.

My first card was Power Color ATI 7870. Blue screened as mentioned above.

Second card was the same model (RMA). It blue screened as the first.

Third card is the same scenario as card 2

Fourth card, they gave me a brand new ATI R9 270. Very nice card, a little faster, still blue screens.

These are the drivers I have installed/uninstalled for all of the cards:
13.9
14.4
14.12 (Omaga)
15.7 (Currently running driver)

I am starting to think that it is actually not the card. Could you guys help me out? Im kind of getting lost into what to do next. Also, can anything be said about changing the bios of my card? I heard people talking about it on the internet and am not sure if it is such a good idea.
View attachment 366442

Any help would be appreciated
Thanks

 

[Arc]

With the R9 200 series card, the 14.12 works good.

There are versatile different bugchecks, which may indicate to a memory related issue. Latest four dumps .....

BugCheck 3B, {c0000005, fffff8801122d3fa, fffff8800a24a750, 0}

Probably caused by : atikmdag.sys ( atikmdag+1353fa )

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

BugCheck D1, {0, 2, 1, fffff8800452d02e}

Probably caused by : hidusb.sys ( hidusb!HumReadReport+13f )

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

BugCheck 109, {a3a039d89746b7e0, 0, df7e17cbe24a89bd, 101}

Probably caused by : Unknown_Image ( ANALYSIS_INCONCLUSIVE )

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

BugCheck 3D, {fffff880030afc80, 0, 0, fffff880122c1137}

Probably caused by : dxgkrnl.sys ( dxgkrnl!DxgNotifyInterruptCB+7b )

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

Test your RAM modules for possible errors.

How to Test and Diagnose RAM Issues with Memtest86+
​

Run memtest86+ for at least 8 consecutive passes.

If it start showing errors/red lines, stop testing. A single error is enough to determine that something is going bad there.

Let us know the result. A camera snap of the memtest86+ window before closing the program would be good.

 

[computergeekguy]

I have run the Memtest 86 program. I have so far managed to get in 6 passes with no errors so far. I am not expecting to see any in the next two passes however so I might start looking at my power supply as the next possible problem. I just bought a new EVGA 600 watt and it arrived not to long ago. As soon as Memtest is done I think I will install it and see if that makes a difference. If that doesn't work what could it possibly be? Is there a way to determine if a program might be conflicting with my card?

Could it possibly be a problem with new video card and older motherboard? I think that I have the latest bios so my mobo is as good as its going to get. Also I think that my card is PCI-e version 3 and my mobo is version 2 so might that conflict? (I heard it wasn't supposed too)

Here is the picture of memtest. View attachment 367100

 

[Arc]

Any memory test less than 8 passes is just waste of time.

 

[computergeekguy]

I got it up to 10 passes. Here is the picture, it still is the same result.

View attachment 367115

 

[Arc]

Try two other hardware tests.

Stress test the Graphics Card using Furmark.

Video Card - Stress Test with Furmark
​

Take a screenshot of the furmark window before closing it. Upload the screenshot for us. Also let us know if you have experienced any crash/BSOD and/or artifacts during the test.

Stress test the CPU.

Hardware - Stress Test With Prime95
​

It saves the result as a .txt file in the prime95's folder. Upload the file for us.

As the range of the bugchecks are very wide, we need to make it sure at first that the main components of the hardware is not failing.

 

[computergeekguy]

I just ran furmark. I haven't gotten to the prime 95 test yet.
I ran furmark for and hour and a half. Here are some specs that I noticed about the test.

Core temp stayed mostly on 75 or lower. Highest was 77

VRM temp got up to 95 (ran it at 95 degrees for about and hour to see if it would crash) Then I placed a fan and cooled it to 82 for the last half hour.

No artifacts or blue screens. In fact my card seems to be more stable when stress testing then when gaming. :huh:

I got a picture plus some other monitoring programs if interested.

View attachment 367208

I will get to melting my CPU tomorrow. icnic:

Thanks

 

[Arc]

Actually, among lot of other bugchecks there is a stop 0x109. It is a mystery bugcheck, and it claims two things to resolve it. Either the BIOS update of a clean reinstall.

As far as I am seeing, there is no further opportunity for you to update the BIOS. Though the BIOS is old itself, still it is the latest, Version 2104.

BiosVersion = 2104   
BiosReleaseDate = 08/13/2010

In this situation I think Prime95 will also come back clean, and most probably you will opt for a clean reinstall.

If you think that you can take this way, follow this guide: http://www.sevenforums.com/tutorials/219487-clean-reinstall-factory-oem-windows-7-a.html and use the drivers via windows update.

Hopefully that will give the stability there.

 

[computergeekguy]

Update: Prime 95 failed my CPU. I was going to send the log but I don't have access to my computer at the moment. (only my laptop) I ran both High FFT and blend and both failed all 6 cores in 30 minutes to an hour. (About the same time as my games).
I think it is important to mention that my CPU never got above 41c on the stock cooler so heat is not a problem. I have a slight hunch that I might know what is going on with my CPU. Back about 4 years ago, I was playing around with Over clocking. I never went through with it as I didn't want to hurt anything as I was an noob. I thought that I returned all settings to their default but I my not have.

Could I look into boosting the voltage of the CPU and leaving the clock settings alone as prime 95 indicates to me improperly set clock settings.

Could you help me out any with settings? Can I just adjust the voltage or do I need to adjust other settings as well?

Thanks

Edit: I think the highest a 1055T can safely go voltage wise it 1.42-1.47 V
I think I am currently set at 1.28 V I may raise it somewhere in between like 1.3V and see what happens

 

[Arc]

As you are getting BSODs, you should stop overclocking and run all the hardware components like CPU, GPU and RAM to their default settings. Also set the BIOS to default, too.

How To Clear CMOS (Reset BIOS)​

If it BSODs even with the default hardware settings, run prime95 again. Let us know if it yields the same result.

Then we will have two places remained to look at.

 

[computergeekguy]

I reset the bios and am still having troubles. All my cores failed (except two who I stopped before they failed) in the same way as before via prime 95. What are the two last suggestions?
I will upload the logs for you since I am at my computer again.
View attachment 367570

View attachment 367571

 

[Arc]

What are the two last suggestions?

One is the heat-dust combination; and the other is the power supply.

Make it sure that the computer is not having a dust accumulation inside it. Dismantle it, clean the inside. Clean the air vents and fan grills. Then check the system heat. Is the computer hot? Report us the heat of the computer after a couple of hours of your normal usage. Upload a screenshot of the summery tab of Speccy. Alternatively, you can publish a Speccy snapshot too: http://www.sevenforums.com/tutorials/311593-speccy-publish-snapshot-your-system-specs.html .

Check if the Power Supply Unit (PSU) supplying adequate power to the computer or not.

eXtreme Power Supply Calculator - The only power supply calculator trusted by PSU manufacturers and computer enthusiasts
​

Also let us know the wattage of the installed PSU.

These two will lead us to the conclusion, I hope.

Though, the logs are not reporting any failure.

 

[computergeekguy]

The power supply calculator suggests a 394w psu. I currently have a 600w EVGA psu.
I also feel that I should mention that I have a watt meter that mesures my computer pulling 460 watts at full load so that is my actual usage.
Amazon.com: EVGA 600 B1 80+ BRONZE, 600W 3 Year Warranty Power Supply 100-B1-0600-KR: Electronics.

This is speccy snapshot at normal use:
http://speccy.piriform.com/results/iwYOJ25AyDGPEgSZSfdS7EA

This is speccy snapshot running my pc at max power:

http://speccy.piriform.com/results/JZa3sJAWSlkXbzEgoJT2JRL

I keep my computer well cleaned. There are no dust bunnies as I blow my case out regularly.
The hottest thing on speccy was my motherboard surprisingly (not including video card).

A temp not shown by speccy was the video card ram which runs at 85c at full load.
Thanks

 

[Arc]

The PSU is sufficient, and the temps are totally under control.

In this situation, if the CPU fails, we cannot shift the responsibility to anything else, unfortunately.

 

[computergeekguy]

Thats understandable. I have one more thing I want to try before I write my CPU of as a dud. My mobo does not have the bios to make it compatible with the latest FX series CPU's. I would most likely need to replace it with another Phenom II. In that case I would have to buy it used from someone.

My OS is OEM. Will changing out the CPU give me activation trouble of any sort?
Thanks

 

[Arc]

My OS is OEM. Will changing out the CPU give me activation trouble of any sort?

Changing the CPU will not cause any issue, but changing the motherboard will (activation by phone will be needed in that case).