[Solved] W10 Constant Random Restarts & Occasional BSOD - New Build

T

The Ugly Fugly

n00b
Joined
Jul 26, 2016
Messages
4
Hi All,

I recently put together a new rig and the specs are as follows.

Specs:
  • CPU - i7 6850k
  • Motherboard - ASUS Strix X99 (Latest Bios 0801)
  • Graphics - EVGA 1080 GTX
  • RAM - Corsair DDR4 3334MHz 32GB (CMD32GX4M4B3333C16)
  • SSD - Samsung 950 Pro 512GB
  • PSU - Corsair AX1200i (1200W)

Issue:
I've been coming across issues from the get go with Windows 10. At first I was experiencing a lot of BSOD issues which to me appeared to either be driver or memory related. This was even after updating to the latest drivers from Nvidia for graphics, latest drivers from ASUS for chipset, audio, ethernet & WiFi and Samsung NVMe driver for the 950 Pro.
  • Page Fault In Nonpaged Area
  • Pool Corruption In File Area
  • Driver Overran Stack Buffer
  • IRQL Not Less or Equal

So I decided to do another clean install and try and tackle it step by step. Now I'm finding that after a clean install and letting windows install all the latest updates (prior to installing any software or driver updates) Windows 10 will stutter a bit and then randomly restart. It will constantly do this, and only this morning I'm finding it is going to a blank screen after Windows 10 login.

Attempted:
So far I've tried the following:
  • Memtest86 on each RAM Stick (Checked out on 4 passes)
  • Clean install Windows 10, multiple times (and feels like its getting worse and sluggier each time)
  • Latest Driver Install from ASUS & Nvidia
  • BIOS Settings Default (i.e. turned off XMP Profiles, Re-enabled CSM)

I'm out of ideas and no step closer to figuring out if its hardware issue or software/Windows 10 related. I can rule out RAM from the memtest. I'm hoping to dump the 1080 GTX into my partner's computer tonight to see if its the card issue. I'll probably try another clean install of W10 on an older SSD tonight. If that fails I'll go back to W7 and see if I have the same issue. Not sure how to rule out Motherboard, CPU and PSU issues though.

If anyone has any suggestions or leads it'd be much appreciated and thanks in advance.

---------------

Issue is solved. In the end it turned out the RAM was faulty. After RMA'ing to Corsair and receiving my replacement RAM the PC has been running smoothinly.

Thanks all again for the suggestions and assistance.
 
Last edited:
If you can login or transfer the existing C:\Windows\minidump & memory.dmp. You can install the Debugging Tools on that pc, or I think another pc, and run windbg.exe to find the cause.
You can use this to install just the debugging tools Windows 10 SDK - Windows app development (it installs to Program Files (x86)\Windows Kits\10\Debuggers\x64\windbg.exe)
 
What speed are you running you RAM at? Run it at default speeds and see if it crashes.

Edit: Default speed = 2133, not for the 3333 OC speed.
 
Last edited:
Yeah considering this is a new rig (not working pulls), I'd rule out the hardware first and foremost.
Let us know how it goes with less aggressive RAM settings.
 
If you find that the ram is ok after running the test as they mention above, you will want to try your cpu next. In the last year alone, I have had one 6700k bad of my own, one bad that I just replaced Intel warranty for a customer and someone here had the same issues, a bad memory controller on the cpu. Basically, if you find the issues not to be the other pieces of hardware, the only thing left would be the cpu. I would get it replaced asap if you are within the window of opportunity of the place you bought it from.
 
Thanks for the suggestions so far, I really appreciate it.

-PK- said:
If you can login or transfer the existing C:\Windows\minidump & memory.dmp. You can install the Debugging Tools on that pc, or I think another pc, and run windbg.exe to find the cause.
You can use this to install just the debugging tools Windows 10 SDK - Windows app development (it installs to Program Files (x86)\Windows Kits\10\Debuggers\x64\windbg.exe)
Click to expand...
Unfortunately I forgot to back up the DMP files prior to a clean install so I don't have a copy of them anymore. But I'll definitely keep this in mind for when and if the next BSOD occurs. At the moment it's just random restarts either after I log into windows or not long after, like 5-10min.


Ocellaris said:
What speed are you running you RAM at? Run it at default speeds and see if it crashes.

Edit: Default speed = 2133, not for the 3333 OC speed.
Click to expand...

heatlesssun said:
I'm thinking along similar lines. I'd even try running it at a lower speed or upping the voltage.
Click to expand...

Mav451 said:
Yeah considering this is a new rig (not working pulls), I'd rule out the hardware first and foremost.
Let us know how it goes with less aggressive RAM settings.
Click to expand...
When I was getting the BSOD and had all drivers up to date it was running at 3333MHz using the XMP profile. Ever since I've clean installed and just Windows Updates it's running at the standard profile of 2133MHz. Which is still resulting in the random restarts either straight after Windows login or not long after it 5-10min. I haven't thought about down clocking it further, what would you recommend in that case?


ManofGod said:
If you find that the ram is ok after running the test as they mention above, you will want to try your cpu next. In the last year alone, I have had one 6700k bad of my own, one bad that I just replaced Intel warranty for a customer and someone here had the same issues, a bad memory controller on the cpu. Basically, if you find the issues not to be the other pieces of hardware, the only thing left would be the cpu. I would get it replaced asap if you are within the window of opportunity of the place you bought it from.
Click to expand...
In this case, how would I rule out/test if its a CPU or a Mobo issue. The reason I ask this is because I dont have another 2011-v3 socket mobo to switch out the CPU to do an isolated test on.
 
If you're down to 2133 and memtest is working I'd check memory off the list. CPU seating would be the next thing I'd look at.
 
1. 2133Mhz is absurdly safe haha - that said, even slow speeds aren't a guarantee that RAM isn't the culprit. But it's been a very long time since I've had bad RAM.
Any possibility you can get an entirely new set of DDR4, even just a pair, to test against? This would make it easier to rule out your own 32GB kit, and then certain DIMMs (if it comes down to it).

2. Moreover, it sounds like these are idle BSODs - i.e. no serious PSU loading or GPU loading.
This is what makes me think it's more of your platform (RAM/Mobo/CPU).

3. Well to answer your last question - that's what makes a straight CPU-exchange more ideal.
Because if the second CPU also exhibits problems, at least you've ruled out the CPU. I'd do this after the RAM exchange though - that's obviously quicker/easier to do than a CPU exchange.

4. The last call would be a mobo replacement - which is obviously not ideal...
 
heatlesssun said:
If you're down to 2133 and memtest is working I'd check memory off the list. CPU seating would be the next thing I'd look at.
Click to expand...

Mav451 said:
1. 2133Mhz is absurdly safe haha - that said, even slow speeds aren't a guarantee that RAM isn't the culprit. But it's been a very long time since I've had bad RAM.
Any possibility you can get an entirely new set of DDR4, even just a pair, to test against? This would make it easier to rule out your own 32GB kit, and then certain DIMMs (if it comes down to it).

2. Moreover, it sounds like these are idle BSODs - i.e. no serious PSU loading or GPU loading.
This is what makes me think it's more of your platform (RAM/Mobo/CPU).

3. Well to answer your last question - that's what makes a straight CPU-exchange more ideal.
Because if the second CPU also exhibits problems, at least you've ruled out the CPU. I'd do this after the RAM exchange though - that's obviously quicker/easier to do than a CPU exchange.

4. The last call would be a mobo replacement - which is obviously not ideal...
Click to expand...
Haha, yeah thats what I initially thought, that the 2133MHz would be somewhat safe.

But what ManofGod mentioned had me thinking. Which had me checking up how could I actually test the CPU to rule it out. I came across Intel's Processor Diagnostic Tool. Now I don't know how well/legit it is in terms of interpreting its results. But it confirmed what ManofGod mentioned. A memory controller on the CPU (the IMC, Integrarted Memory Controller test failed). For now I'm going through the pain of trying to get a replacement from the store I bought the processor from. Pain because its actually been sitting there for a few weeks whilst I tried to source a 1080 before doing a full test.

Hopefully a CPU replacement will solve everything. Otherwise I guess like you mentioned Mav451, the last but not ideal thing to check would be the mobo.

Thanks all for the help so far, will touch base again as soon as I'm able to get a replacement which knowing this store may be up to a week or two from past experiences.
 

Attachments

  • IMC Fail.jpg
    IMC Fail.jpg
    169.5 KB · Views: 45
Try running with only 16GB.

I saw something some place about memory controller or chipset issues with a full 32GB.
Sorry I can't be more specific.

ETA: Ooops... missed the post on the IMC fail.

Try running the CPU test again with 16GB for the fun of it though.

.
 
Thanks all again.

In the end it wasn't a CPU IMC issue. It was as Mav451 mentioned - despite passing memtest and running at safer speeds it doesn't necessarily guarantee that the RAM isn't the culprit.

After still experiencing issues on receiving my replacement CPU, I ended up RMA'ing my RAM which I got back a few days ago.

So far the PC has been running really well at both the stock 2133MHz and 3333MHz XMP profile settings.
 
You must log in or register to reply here.
Back
Top