BSOD during install/uninstall of AMD Crimson 16.5.3 display driver, even w/ DDU in safe mode. Possible graphics HW issue??

a_dumb_noob2

Reputable
Oct 30, 2015
20
0
4,510
Hello everyone at Tom'sHardware. As the community of this site has been so helpful in the past, I thought to ask you guys first when I started to become faced with this issue. My setup is as follows:

Windows 10
i7-4790k, OC'd to 4.5Ghz
Noctua NH-D14
ASUS z97 Deluxe USB 3.1 Motherboard
Kingston Fury HyperX 1866Mhz & CL10 RAM, OC'd to 2000Mhz w/ timings increased to CL11
MSI Gaming R9 390X 8GB VRAM, OC'd to 1120Mhz Core/1550Mhz VRAM
Corsair CS850M 850W 80+ Bronze PSU
Two case fans, 1 intake and 1 exhaust (w/ Noctua CPU cooler aiding exhaust)

So my situation began tonight when I tried to update my drivers to AMD's 16.5.3 Crimson driver package. Upon uninstalling using the AMD uninstaller, my computer simply froze. Upon restarting, I found out that my computer had in fact uninstalled all the drivers, so I proceeded with my uninstall by using ccleaner to clean my registry of errors, and then restarted my computer in safe mode to finish the uninstall process using Display Driver Uninstaller. To be clear, I used all the recommended settings and options in DDU to uninstall AMD drivers and simply pressed "Clean and Restart". After the program seemed to finish cleaning up all the remnants of the driver, the computer once again froze when the program started to "scan for new adapter" or something similar (I can't recall exact wording). However, it is always at this point where the computer freezes. In fact, I only found out that the computer was blue screening after the 5th or 6th try when the computer froze and then displayed half of a blue screen. As far as I understand, BSOD during safe mode are a bad sign and typical of HW issues. Could it be possible that my GPU is defective? It's only a year old and I have never overclocked it very hard. In fact, before this week, I didn't even overclock it at all, finding that the stock speeds were absolutely fine for me. Regardless, I have had the card working before, and I had previous versions of the Crimson drivers installed to run it. That being said, the installation of that driver (which was the first after a fresh install of Windows 10), also presented me with no end of issues itself. Though, I did eventually get everything to work so Im sort of lost for an explanation.
My only next obvious course of action beyond starting this thread on the message board, is to see if the computer left any memory dumps upon crashing so that I can run the debugger to see what might be going on. I will let you all know upon doing so but I would still like to ask if anyone has any potential fixes/ideas/explanations/etc. for me in the meantime while I try to sort things out. Thanks for your time guys and gals. I appreciate any and all feedback (as long as it's civil). I await your responses. :S

 
Okay, I'll give it a shot. I'm going to try the 16.3.2 drivers and report back with the results.
Also, I don't seem to have any system memory dumps. I do however have a couple crash dumps labelled "rundll32.exe.####.dmp", "igoproxy64.exe.####.dmp", and "RadeonSettings.exe.####.dmp" (with actual digits replacing the '#' signs). Would these be of any use at all?
 
Okay, so I just tried to revert back to the 16.1 beta drivers which was what I had installed initially way back when on my fresh Win 10 install and unfortunately, I still have issues (i.e. it's failing to install properly). Luckily, there weren't any BSODs though it failed to complete (but the new driver doesn't BSOD always either but also still fails). I'm still lost for a solution. Anyone have any suggestions?

I will try the 16.3.2 drivers that AMD offers on the website under the notes of the 16.5.3 driver.
 
I doubt those will be useful in this case. A kernel dump would be better. Make sure you've got dumps enabled, sometimes they aren't by default.
From "This PC" (the new My Computer), right click the white space and click properties. Click advanced system settings on the left panel, then the startup and recovery button. Make sure the drop down box is on kernel memory dump or larger.
 
So after downloading AMD Crimson 16.3.2, I went ahead and tried to uninstall everything from 16.1 Beta, and in safe mode while using Display Driver Uninstaller, I got another freeze as usual and at the same spot. This time however, I got a full BSOD with the error code "CLOCK_WATCHDOG_TIMEOUT", which I would have to assume means that my CPU overclock is totally unstable. Is this right? If so, I'll drop the overclock for now and see if that doesn't fix the issue.

As for memory dumps on crashes, I believe I have it set to create small kernel dump files.
 
The best step for any troubleshooting would be to undo any overclock you have. This includes CPU, GPU, and RAM. Any one of those could trigger a BSOD. Post up the latest dump files and I'll see if I can narrow it down. (Usually C:\Windows\Minidump\ and C:\Windows\MEMORY.dmp)
 


Okay thanks. I'll ditch the overclocks at the moment and I'll try installing the drivers again. In the meantime though, I don't have any dump files so I honestly hope that I can get everything to work going forward so we don't need to analyse crash dumps but I've got everything set up to make dump files if the PC BSODs again.

Btw, I just want to thank you for your help so far. I really appreciate it. The Tom'sHardware community is some of the best people around and you guys never seem to fail me :)
 
Okay so gotten rid of the overclocks on all my hardware but upon trying to install any graphics driver, I still get crashing. Sometimes a BSOD, sometimes just a frozen computer.
 
Okay, upon trying to install driver 16.3.2, my computer froze. When I hard reset it, I got two BSODs immediately with the error "PAGE_FAULT_IN_NON_PAGED_AREA" with one showing atikmdag.sys as the relevant file. Now I'm autobooted by the system into the Automatic Repair.

Any ideas?
 
More bad news. I cannot seem to get into either Windows 10 proper or safe mode. Upon booting every time, I receive the "PAGE_FAULT_IN_UNPAGED_AREA". Considering I've read elsewhere on the web that this could be caused by faulty memory or bad sectors on the HDD, I swapped my memory sticks around on the motherboard and also ran the chkdsk command in command prompt. Strangely, when I tried to set the bcdedit to boot to safe mode, I couldn't restart the computer without exiting command prompt first because my "shutdown" commands arent working (likely caused by a missing shutdown.exe file in the PATH). Anyways, as you can see here though, my computer is basically a freaking mess at this point and I have absolutely no friggin clue what to do. The memory seems to be fine and the chkdsk command didn't seem to pull any problems up. And since this seemed to happen because of a driver install, I'm almost certain this is graphics driver related. Goddamn this is a pain....