Jump to content

Welcome to Geeks to Go - Register now for FREE

Geeks To Go is a helpful hub, where thousands of volunteer geeks quickly serve friendly answers and support. Check out the forums and get free advice from the experts. Register now to gain access to all of our features, it's FREE and only takes one minute. Once registered and logged in, you will be able to create topics, post replies to existing threads, give reputation to your fellow members, get your own private messenger, post status updates, manage your profile and so much more.

Create Account How it Works
Photo

Unidentifiable GPU death or driver issue (Solved)


  • Please log in to reply

#1
Oniketsoku

Oniketsoku

    Member

  • Member
  • PipPipPip
  • 340 posts

Hello G2G
 

My rig has been running pretty okay for quite some time, but in the past 4-6 weeks the situation has been slowly degrading to a grinding halt as of last night when SHTF finally. Performance started to drop in applications and my PC would crash when running demanding games so I stopped playing them. Last night, I couldn't even play League of Legends on the bare minimum settings for 5 minutes without a crash, half the times it'd recover with a message from the systray saying nVidia driver 3.35 such and such has recovered, other half I'd just get a forced restart. I tried updating the GPU driver via windows update, but there seemed to be literally zero influence or change afterwards.

This morning I cleaned my PC as quite a bit of dust had built up and thought maybe that was the culprit. Turns out I can't even boot up any more, just get a loop of restarts. First time after getting everything hooked back up I got a BSOD but I can't recreate the blue screen to record the error. I can only get things functional by running in safe mode with networking (which is where I am posting to you from). It sounds like a driver issue but a slow degredation over time suggests a hardware issue, right? I honestly don't know. I've been mostly clean on my MWB scans for over a year but I guess I couldn't totally rule out malware as a problem since I don't scan ultra deep or anything. I know the first step I have to do is isolate and identify the issue but I'm really not too sure where to start with this and there's a ton of misinformation on the net. Please help!

Anyways, the graphics card in question is a Zotac GTX 560 ti. I'll compile a list of stuff I've done/researched prior to posting here:

I think that's everything. I'm considering doing the whole msconfig/disable all services thing but am not sure if it's relevant to my problem. Let me know what you think and how I should start trying to fix this. I haven't tried doing a normal boot with onboard GFX and removing the card, but I'm 99.9999% sure it would work. I also considered putting on some new thermal while cleaning but temperatures have never really been an issue. If I make any changes or try something else I'll make a note of it in the OP.

Thank you in advance.

EDIT: Tried to run DDU to remove all GPU drivers on the system and install the newest one fresh and this was the result http://puu.sh/dwo9V/520f9973bd.png

 

EDIT 2: Reseated the GPU, ran sfc /scannow, ran ComboFix, ran FRST, then ran DDU again in safe mode and it worked - deleted all AMD and nVidia drivers, restarted PC with internet off & installed newest driver, here was the result after the final restart and opening the client to test a game (not even in the game yet, just opened the launcher): http://puu.sh/dwv6M/bb6eb20896.png  From someone ignorant like me on the outside looking in, I think it has to be a conflict with some sort of other drvier, windows update, software, or failing hardware. Maybe the card is failing to use/apply the driver properly and that type of error code is the only way it knows how to show something is wrong? IDK. I'll try disabling the services from msconfig tomorrow morning since that is more suspect now. I'll look around for some sort of video card diagnostic in the meantime 

EDIT 3: Ok, I found a decent diagnostic named GPUZ. Here are the results after running stuff randomly for a few minutes: http://puu.sh/dwxxL/b0322f66d1.jpg the screen blacked out and I got the Windows Kernel Error Driver systray message after the fan speed hit 0 and the GPU load spiked to over 99%. I attached a .txt of the raw data Attached File  gpuz log number one.txt   76.87KB   69 downloads if anyone is interested, I have found the problem in action but no idea how to interpret it yet (open the log with notepad otherwise it's ineligible). In the log you see the GPU load spike REALLY FAST over a few seconds straight to 99 then go back to normal. It'd be really easy to blame it on the PSU or temperatures getting too high but you can see that temps are stable and there are no voltage drops. I am officially stumped! 


Edited by Oniketsoku, 15 December 2014 - 10:27 PM.

  • 0

Advertisements


#2
iammykyl

iammykyl

    Tech Staff

  • Technician
  • 6,763 posts

Welcome back Oniketsoku.

 

To get your system specs< please use Speccy and paste the URL in your next reply, > http://www.bleepingc...t-using-speccy/

Also a look at disks, Please copy and paste, diskmgmt.msc in the Run box and include a screenshot in your next reply.   When the window opens, you may have to drag the right side and bottom out, so all contents can be seen. 


  • 0

#3
Oniketsoku

Oniketsoku

    Member

  • Topic Starter
  • Member
  • PipPipPip
  • 340 posts

Hiyo, thanks mykyl

http://speccy.pirifo...tlrfMqeEiyOoshV that's pretty nifty thanks for the protip

 

052ab8d41a.png

Per your request

 


  • 0

#4
iammykyl

iammykyl

    Tech Staff

  • Technician
  • 6,763 posts

Thanks for the info.

I believe much of your slowdown, poor performance, freezing, is lack of free space on the hard drive.   If a drive gets too full then you get Boot problems.   A drive needs space for Temps, Cache, paging, etc.   Lots of opinions on this, but I keep 10% free on a SSD and 20% on a mechanical drive.   If a drive is fragmented, Data is placed, a bit hear and a bit there, all over the drive, so free space should be contagious so Data is written in blocks.   As the drive fills you start to access the inner area of the platters, the slowest part.

 

Let's start with a cleanup to see if things improve, then we can look at other issues.   The one program i do not use or recomend is CCcleaner, up to you, but it does cause lots of problems by cleaning the registry,

Boot to safe mode with Networking and follow this guide, (do not clean out service pack files, etc. or clean the WinSxS Folder, mistakes can easily be made. 

>  http://www.online-te...ard-disk-space/

When finished, defrag using Auslogics Disk Defrag 5.1.0.0 Download from this site, Warning, during installation, untick the box for installing Boost Speed. > http://www.majorgeek...isk_defrag.html


  • 0

#5
Oniketsoku

Oniketsoku

    Member

  • Topic Starter
  • Member
  • PipPipPip
  • 340 posts

Okay

be9c8bfd2e.png

I'm not sure this will aid with the GPU but it was definitely beneficial to my system and being proactive is good so I'll collect a new sample of data and see if stuff still crashes


Edited by Oniketsoku, 16 December 2014 - 12:52 PM.

  • 0

#6
iammykyl

iammykyl

    Tech Staff

  • Technician
  • 6,763 posts

 

 

In the log you see the GPU load spike REALLY FAST over a few seconds straight to 99 then go back to normal.

My understanding is that it is normal behaviour due to power saving,   When not under load, all setting drop, when you, say, load a new level, view/screen or intense action displayed like explosions, fast action with lots of detail, all setting increase to perform the task/s

 

You have performed the steps I would use without a fix so see if we can get some info.

In the run box, copy and paste eventvwr.exe   How to use,http://www.howtogeek...-a-pro/lesson3/

Run the Realiabilty Monitor, How to use, > http://www.makeuseof...bility-monitor/

Download HWifor64, upload a screenshot of the volts/temps/fan,  > http://www.majorgeek...s/hwinfo64.html Warning, watch out for other software when installing. 


  • 0

#7
Oniketsoku

Oniketsoku

    Member

  • Topic Starter
  • Member
  • PipPipPip
  • 340 posts

Running those next steps now. Just thought I'd post a follow-up. Everything seemed to be running okay for about an hour and then it started crashing again. The same problem persists but the overall quality of my machine has improved.


EDIT:

These are some pretty useful tools I didn't know about. Thanks for sharing.

be9fd9ec2a.png

Looks like the origin of this problem was all the way back in August. Was there something specific you wanted me to look up with Event Viewer though?

c6fcb48737.png

And here's the HWifor64 stuff

                                     Current        Minimum   Maximum      Average 
7734e8b384.png

It's not a real large amount of time to collect data so I'll post another image after it has been collecting for several hours. May be worth noting that last night it had a crash simply while watching a 2 minute trailer on Youtube last night. Granted it was 1080p and 60fps footage of some intense stuff but I don't know if there's any actual rendering involved with just watching a video... weird
 


Edited by Oniketsoku, 17 December 2014 - 09:22 AM.

  • 0

#8
Oniketsoku

Oniketsoku

    Member

  • Topic Starter
  • Member
  • PipPipPip
  • 340 posts

I actually didn't have any crashes today but here's the day's results
 14a094a054.png


Edited by Oniketsoku, 17 December 2014 - 05:31 PM.

  • 0

#9
iammykyl

iammykyl

    Tech Staff

  • Technician
  • 6,763 posts

Sorry for the slow response, a bit busy at the moment.

The event view shows no errors, good.  Just clear the logs then inspect (just once) at the end of the day, report any errors.

 

 

 

It's not a real large amount of time to collect data so I'll post another image after it has been collecting for several hours.

OK.   Look at the details for any new errors, see if there is a solution.

 

The GPU reading look good.

 

Start HWinfo, run a CPU stress test, keep HWinfo on top, make sure the CPU info is displayed..   After 2/3 minutes, take a screenshot, please include in your next reply. 

***********************

Please look at the recent installed Windows Updates. search for  the update named KB3004394  ???  (may be the cause of your issues)

Then search for KB3024777  ???   (removes the problem update and fixes issues)


Edited by iammykyl, 18 December 2014 - 05:09 PM.
Further instructions added.

  • 0

#10
Oniketsoku

Oniketsoku

    Member

  • Topic Starter
  • Member
  • PipPipPip
  • 340 posts

Sorry for disappearing, have been busy with a multitude of things. Unfortunately, it definitely seems like things are getting worse. Getting a lot more crashes and blackouts when not even gaming. I've actually had two in the middle of writing this, lol. Also, got a new error last night and was unable to run the PC unless in safe mode again. I was hoping it was that windows update as well but unfortunately that was one of the first things I checked back in the OP

cab1d45fd7.png
Which is pretty interesting and led me to finding this thread: http://www.techsuppo...sys-604177.html seems like he was having the same problem and the only thing that worked was replacing the mobo. I am thinking of purchasing one to see if it's the fix I need and if it doesn't work, just returning the board. What are your thoughts?

Anyways, I will do and upload the CPU stress test you requested in the meantime. Was there a particular way you recommend going about it before I search for one?


  • 0

Advertisements


#11
Oniketsoku

Oniketsoku

    Member

  • Topic Starter
  • Member
  • PipPipPip
  • 340 posts

Here's the result of the test 58e61b92ea.png

I left it running for about 2 minutes, forgot to reset the timer on HWinfo for the first 30sec. No crashes, seemed to run fine


Edited by Oniketsoku, 21 December 2014 - 10:25 PM.

  • 0

#12
iammykyl

iammykyl

    Tech Staff

  • Technician
  • 6,763 posts

That was a good result, it also puts a  high demand on the PSU, so all good. 

A replacement board  sounds like a plan.   How much would you loose if you had to return, restocking fee?

Pleas check if you have Asus EPU engine installed, if yew, configure, not to start with Windows, then stop it.

 

In safe mode with networking.

Go to > http://www.asus.com/...pDesk_Download/

Download and save to the desktop, ASUS GPU NOS V1.00.17 for Windows XP 32bit/XP 64bit/Vista 32bit/Vista 64bit/7 32bit/7 64bit then install.

When finished, restart and enter the BIOS.

Go, > Advanced Menu, > CPU Configuration, > Cool N Quiet, > Disable,

Go, > Internal Graphics > Primary Graphics Controller, > Select GFX0.

Go, > Exit, > Press F10, > Confirm.   

Test.   if no good.

 

Create a manual Restore Point, name it igfxCUIService.exe set manual

In the Run Box, > copy and paste Services.msc

Look for gfxCUIService.exe (if present) > right click > click Properties, > Change the Startup Type to Manual.

If the service is running, > click Stop, > click Apply, click OK. > close all open windows.

Restart the computer.

Test.

 

 

 

 

Sti 


  • 0

#13
Oniketsoku

Oniketsoku

    Member

  • Topic Starter
  • Member
  • PipPipPip
  • 340 posts

Got this when trying to install GPU_NOS while in safe mode w/nw
6f35c300a5.png



I decided to go with this board http://www.newegg.co...N82E16813128514
I cross-checked my hardware to make sure it was compatible and it looks ok to me. The board is AM3+ but the CPU itself is AM3, yet that doesn't seem to be an issue after some googling

It seemed like a good idea at the time because of the 30 day free trial for the newegg premier thing with free 2-3day shipping, free returns, and no restocking fee

In the meantime I'll try the gfx CUI service fix you recommended and post back after it's done, but no luck with the first one

EDIT: As of now I can't even to get it to boot so only safe mode works, but it doesn't even want to bring up safe mode as an option. I'll let it power off until tomorrow that usually fixed it before (which yells hardware problem at me personally)


Edited by Oniketsoku, 22 December 2014 - 07:22 PM.

  • 0

#14
iammykyl

iammykyl

    Tech Staff

  • Technician
  • 6,763 posts

Strange that would not install.   Happen it needs a chipset upgrade first.   Create a restore point, then ry going back and download/install AMD Chipset Driver V1.3.2.54, test, then try the GPU_NOS again. 

Yes, your CPU is on the supported list. > http://www.gigabyte....p.aspx?pid=4672

 

Yes, try the gfx CUI service fix. 


  • 0

#15
Oniketsoku

Oniketsoku

    Member

  • Topic Starter
  • Member
  • PipPipPip
  • 340 posts

As soon as I try to run the installer it immediately hangs and becomes unresponsive after accepting the license agreement, tried 6 or 7 times

eb08f1e9e3.png

Also, booting in anything other than safe mode is futile now. When it does manage to get into windows normally it won't last more than 2 minutes


UPDATE:

I left it running unresponsive for 10-15 minutes and it popped up with this all of a sudden

a64697ad0e.jpg

Unfortunately view log won't work or bring anything up

Also, I can't find gfx CUI anywhere in my services list. 95e055b5e1.png


Edited by Oniketsoku, 23 December 2014 - 11:01 AM.

  • 0






Similar Topics

0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users

As Featured On:

Microsoft Yahoo BBC MSN PC Magazine Washington Post HP