BSOD on "cold" boot - stable afterwards - Long-term issue

Mentok · 10 Mar 2016

Hello everyone, I'm at my wit's end at the moment.

Recently had to replace motherboard, RAM and CPU in my computer. Purchased an ASUS Maximus VIII GENE Z170 board, an i5-6600k and initially 16GB of DDR4, now 32.

When I boot my computer after having it off overnight, I can get to the login screen, sometimes beyond before it hangs and gives me a BSOD. After the machine reboots, it's stable for as long as I leave it on. I've stressed the system, played games, done everything to try and get another BSOD after the machine is "warmed up" with nothing. I've run Prime95 for 4 hours and completed 2 whole memtests on both the 16 and 32GBs of RAM. I've swapped the RAM, and even replaced the motherboard with an exchanged unit (so everything has been pulled and re-seated at least once).

Sometimes it seems to boot fine and it stays stable. Restarting the machine without a BSOD does not produce a BSOD, the system is fine. There are no OCs on the GPU, CPU or RAM.

Really puzzled and hoping for some help!

Attachment 68678

Ztruker · 10 Mar 2016

Most reccent dumps shows:

MODULE_NAME: nvlddmkm
IMAGE_NAME: nvlddmkm.sys

Code:

Module Name  Display Name           Description            Driver Type   Start Mode State      Status     Accept Stop Accept Pause Paged Pool(bytes) Code(bytes) BSS(bytes) Link Date              Path                                             Init(bytes)
============ ====================== ====================== ============= ========== ========== ========== =========== ============ ================= =========== ========== ====================== ================================================ ===========
nvlddmkm     nvlddmkm               nvlddmkm               Kernel        Manual     Running    OK         TRUE        FALSE        1,466,368         6,668,288   0          3/3/2016 4:15:05 AM    C:\Windows\system32\DRIVERS\nvlddmkm.sys         32,768

This is your Nvidia driver. You have a very recent (3/3/2016) version. Maybe try rolling back to a previous version (or older?). Do it cleanly though. Uninstall the current, reboot and let Windows install whatever it thinks is the right driver. Test, if still problems then try going back a few versions.

An older dump (02/08/16) shows:

MODULE_NAME: USBXHCI
FAULTING_MODULE: fffff800eba80000 USBXHCI

Code:

Module Name  Display Name           Description            Driver Type   Start Mode State      Status     Accept Stop Accept Pause Paged Pool(bytes) Code(bytes) BSS(bytes) Link Date              Path                                             Init(bytes)
============ ====================== ====================== ============= ========== ========== ========== =========== ============ ================= =========== ========== ====================== ================================================ ===========
USBXHCI      USB xHCI Compliant Hos USB xHCI Compliant Hos Kernel        Manual     Running    OK         TRUE        FALSE        53,248            241,664     0          10/29/2015 10:32:00 PM C:\Windows\system32\drivers\USBXHCI.SYS          4,096

AiChargerPlus.sys is also mentioned which belongs to ASUS and some posts I've found about it indicate it has caused problems previously. Don't know if ASUS fixed the problem or not.

A couple of the oldest dumps show:

MODULE_NAME: memory_corruption
IMAGE_NAME: memory_corruption
FOLLOWUP_NAME: memory_corruption

Have you tried testing your system RAM with memtest86+?

Mentok · 10 Mar 2016

Thanks for the speedy reply!

Nvidia Driver
I have used different drivers, as you can likely see in the logs, and it doesn't seem to make a difference which one I am using. I've even had the BSOD when using the Intel integrated graphics.

ASUS USB
The AI charger was installed as part of an "install everything that ASUS puts out for the VIII GENE" experiment, to see if there was some mystery driver I was missing. The BSODs continued.

Memtest 86+
I have run the test on both sets of RAM, and then together already. However I will run them again tonight for a minimum 8 passes, exactly as outlined on the board.

I'll report back when this is complete.

Also, what do you make of this only happening when the machine is physically "cool"? Once the machine is up for a few minutes, it's rock-stable. It's only the first boot after it's been off for a few hours that it may BSOD (and if it doesn't, it's stable). Ever seen anything like that?

Ztruker · 11 Mar 2016

When you get the BSOD from a cold start, have you tried shutting down the machine then powering up and if so does that cause another BSOD.

You replaced then swapped out the mother board, tried different RAM combinations so I think that eliminates those components.
That leaves your CPU and power supply as possible suspects, if it's a hardware problem. Any chance to old CPU would work in the new board as a test?

Did you do a clean install of Windows 10 after the motherboard, RAM and CPU replacements?
New board, what about a BIOS update?
Are you over clocking at all?

Mentok · 11 Mar 2016

I left the memtest86+ running overnight - it only got through 2 passes and a bit. That seems exceptionally slow- see the screenshot. It was also running on only one core. When I tried using multiple cores, it got quite a bit further, but froze up about 35 minutes in on the block moves.

As to your most recent post:
I have tried shutting down and restarting again, but the BSOD doesn't re-occur.
The PSU is an EVGA SuperNOVA 1000 G2, purchased about 1.5 years ago (when I was running some high-draw SLI config). The old CPU is a 3930k, so it won't fit the 1151 socket.

Windows 10 was a clean install when I built the new system. Clean install again when I replaced the new motherboard.
BIOS has been tried at release version, and the last 2 updated versions (as they came out). No problems flashing them.

I haven't applied any OC, and disabled some of the OC-boost the board enables for a "k" series processor. Voltages and clocks are all box-stock.

Ztruker · 11 Mar 2016

What is your hard drive setup?

Have you tried running the HD manufacturers diags to see if there is a problem? Could try running them from a booted CD or flash drive when the computer is cold to get as close to the fail scenario as possible.

7 Free Hard Drive Testing Programs
Hard Drive Diagnostics Tools and Utilities (Storage) - TACKtech Corp.
Bootable Hard Drive Diagnostics

Ztruker · 11 Mar 2016

Might be time to run Driver Verifier and see if that gives us any useful info to work with.
See here: Driver Verifier - Enable and Disable in Windows 10

Take it step by step:

Part One: Run Driver Verifier
Part Two: Configure Driver Verifier Settings and Enable
Part Three: Disable Driver Verifier

axe0 · 12 Mar 2016

I left the memtest86+ running overnight - it only got through 2 passes and a bit. That seems exceptionally slow- see the screenshot. It was also running on only one core. When I tried using multiple cores, it got quite a bit further, but froze up about 35 minutes in on the block moves

Looking at your RAM (32GB) it takes a lot of time to let the test finish completely.
8GB takes about 8-10h, 32GB takes at least 32h and possibly up to 36h, both for 8 complete passes.

Mentok · 13 Mar 2016

Ran Driver Verifier for a couple days- no BSODs to report.

axe0 · 13 Mar 2016

Run MemTest86+ again but this time for at least 8 complete passes.

Diagnostic Test

RAM TEST

Run MemTest86+ to analyse your RAM. MemTest86+ - Test RAM - Windows 10 Forums

Note

MemTest86+ needs to be run for at least 8 complete passes for conclusive results. Set it running before you go to bed and leave it overnight. We're looking for zero errors here. Even a single error will indicate RAM failure.

Addition:
If errors show up you can stop the test, remove all sticks but 1 and test this single stick in each slot for 8 passes or until errors show, switch sticks and repeat.
If errors show up and you see them a lot later, no problem, the errors don't affect the test.