New i2700k Z68 PC w/random weekly BSOD

[fixbsod]

Just got a new computer with the following specs:

i7-2700k running at stock speed + voltages
16 GB 4x4 Corsair Vengeance 1866 (CMZ8GX3M2A1866C9) running at 1600 9-10-9-27-2
Gigabyte Z68XP-UD3 rev1.0 (PCIe 2.0) running latest F8 bios
eVGA Geforce 580 1.5GB card most current non beta WHQL (285.36?)
Windows 7 Ultimate 64-bit w/SP1 and all hotfixes/patches
Corsair HX 1050 Watt PSU
I'm using a Corsair H60 cpu cooler and the case is well ventilated (Corsair 600T w/2x200mm fans)


I am getting a BSOD about once per week of uptime but not for any discernible reason, ie the majority of them have happened while the computer was left torrenting overnight, but with minimal use. The most recent BSOD did happen while playing BioShock 2, although the game was paused in the manu screen for about 5-10 minutes before the BSOD. Uptime does not seem to matter -- the most recent BSOD was on like 1 day of uptime and yet the computer was previously up for over 7 days with no issues -- had to reboot for some Windows update.

I have posted (the lengthy) WinDBG debugged BSOD below, but basically it is an IRQL_NOT_LESS_OR_EQUAL error w/bugcheck 0x0A and filename svchost.exe . I have done numerous MemTest+ 4.20 runs and even reseated/rearranged the RAM and see no issues with the RAM. The machine is stable for ordinary use -- have played 35+ hours of Skyrim with only one softlock (able to go back to Windows and re-load program, so likely just some minor hiccough), 25+ hours BioShock 1 no problem and just the 1 BSOD w/25+ hours of BioShock 2 -- so only 1 BSOD with 100+ game hours. Likely more game hours as this doesn't include some a few HL2 or TF2 hours.

The issue seems to be some driver conflict, however I do not know how to nail it down as the named file is a windows host file and not a specific driver file which would reveal the conflict. The computer is brand spanking new so there were few if any old drivers installed, the OS is a legit W7 64bit w/SP1 and all the updates, and minimal software installed as it is so new. There is no resident virus/malware scanner and sleep settings are all disabled. Nothing is overclocked. The BSODs are not preceded by any erratic behavior (no odd sounds/windows) and I have looked through the event viewer but found no pattern or any issues before the BSOD. If anyone had any suggestions I'm all ears as my next 2 steps are :

1) Installing brand new G.Skill Ripjaws DDR3-1600 4x4GB RAM in lieu of the Corsair.
and then the less desirable
2) Formatting and reinstalling W7
----
Microsoft (R) Windows Debugger Version 6.12.0002.633 AMD64
Copyright (c) Microsoft Corporation. All rights reserved.


Loading Dump File [C:\Windows\MEMORY.DMP]
Kernel Summary Dump File: Only kernel address space is available

Symbol search path is: SRV*c:\symbols*http://msdl.microsoft.com/download/symbols
Executable search path is:
Windows 7 Kernel Version 7601 (Service Pack 1) MP (8 procs) Free x64
Product: WinNt, suite: TerminalServer SingleUserTS
Built by: 7601.17640.amd64fre.win7sp1_gdr.110622-1506
Machine Name:
Kernel base = 0xfffff800`03005000 PsLoadedModuleList = 0xfffff800`0324a670
Debug session time: Tue Jan 24 20:25:10.204 2012 (UTC - 8:00)
System Uptime: 1 days 1:13:26.031
Loading Kernel Symbols
............................................................ ...
............................................................ ....
............................
Loading User Symbols
PEB is paged out (Peb.Ldr = 000007ff`fffd3018). Type ".hh dbgerr001" for details
Loading unloaded module list
.....
************************************************************ *******************
* *
* Bugcheck Analysis *
* *
************************************************************ *******************

Use !analyze -v to get detailed debugging information.

BugCheck A, {48, 2, 1, fffff8000312cfcf}

Probably caused by : memory_corruption ( nt!MiIdentifyPfn+26f )

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

6: kd> ! analyze -v
************************************************************ *******************
* *
* Bugcheck Analysis *
* *
************************************************************ *******************

IRQL_NOT_LESS_OR_EQUAL (a)
An attempt was made to access a pageable (or completely invalid) address at an
interrupt request level (IRQL) that is too high. This is usually
caused by drivers using improper addresses.
If a kernel debugger is available get the stack backtrace.
Arguments:
Arg1: 0000000000000048, memory referenced
Arg2: 0000000000000002, IRQL
Arg3: 0000000000000001, bitfield :
bit 0 : value 0 = read operation, 1 = write operation
bit 3 : value 0 = not an execute operation, 1 = execute operation (only on chips which support this level of status)
Arg4: fffff8000312cfcf, address which referenced memory

Debugging Details:
------------------


WRITE_ADDRESS: 0000000000000048

CURRENT_IRQL: 2

FAULTING_IP:
nt!MiIdentifyPfn+26f
fffff800`0312cfcf f0410fba6e481f lock bts dword ptr [r14+48h],1Fh

DEFAULT_BUCKET_ID: VISTA_DRIVER_FAULT

BUGCHECK_STR: 0xA

PROCESS_NAME: svchost.exe

TRAP_FRAME: fffff880064143a0 -- (.trap 0xfffff880064143a0)
NOTE: The trap frame does not contain all registers.
Some register values may be zeroed or incorrect.
rax=0000000000000001 rbx=0000000000000000 rcx=0200000000000020
rdx=000000000000c150 rsi=0000000000000000 rdi=0000000000000000
rip=fffff8000312cfcf rsp=fffff88006414530 rbp=fffffa800d190930
r8=000000000000c151 r9=0000000000000001 r10=0000000000000042
r11=0000058000000000 r12=0000000000000000 r13=0000000000000000
r14=0000000000000000 r15=0000000000000000
iopl=0 nv up ei pl nz na po nc
nt!MiIdentifyPfn+0x26f:
fffff800`0312cfcf f0410fba6e481f lock bts dword ptr [r14+48h],1Fh ds:00000000`00000048=????????
Resetting default scope

LAST_CONTROL_TRANSFER: from fffff800030811e9 to fffff80003081c40

STACK_TEXT:
fffff880`06414258 fffff800`030811e9 : 00000000`0000000a 00000000`00000048 00000000`00000002 00000000`00000001 : nt!KeBugCheckEx
fffff880`06414260 fffff800`0307fe60 : 00000000`42506650 00000000`00000000 00000000`00000000 02000000`00091529 : nt!KiBugCheckDispatch+0x69
fffff880`064143a0 fffff800`0312cfcf : 00000000`00377ff8 fffffa80`1198a6c0 00000000`42506600 fffff800`0336d73f : nt!KiPageFault+0x260
fffff880`06414530 fffff800`0312dc7b : 00000000`00000000 00000000`00000004 fffffa80`1198a6d8 fffffa80`11989000 : nt!MiIdentifyPfn+0x26f
fffff880`064145d0 fffff800`034927e5 : fffffa80`11989000 fffff880`06414b60 fffff880`064146a8 00000000`00000000 : nt!MmQueryPfnList+0xbb
fffff880`06414610 fffff800`033d54c8 : 00000000`00000006 00000000`00000000 fffffa80`11989000 00000000`00000001 : nt!PfpPfnPrioRequest+0x115
fffff880`06414660 fffff800`0338bbd3 : 00000000`00000000 00000000`00000000 00000000`64496d4d fffffa80`64496d01 : nt! ?? ::NNGAKEGL::`string'+0x4810d
fffff880`064146f0 fffff800`0338c449 : 00000000`0187bd68 00000000`00000000 00000000`0187bdc0 00000000`0187bd88 : nt!ExpQuerySystemInformation+0x1193
fffff880`06414aa0 fffff800`03080ed3 : 00000000`00000001 fffff880`06414b60 00000000`02ddf7e0 00000000`02f472e0 : nt!NtQuerySystemInformation+0x4d
fffff880`06414ae0 00000000`777b167a : 00000000`00000000 00000000`00000000 00000000`00000000 00000000`00000000 : nt!KiSystemServiceCopyEnd+0x13
00000000`0187bc98 00000000`00000000 : 00000000`00000000 00000000`00000000 00000000`00000000 00000000`00000000 : 0x777b167a


STACK_COMMAND: kb

FOLLOWUP_IP:
nt!MiIdentifyPfn+26f
fffff800`0312cfcf f0410fba6e481f lock bts dword ptr [r14+48h],1Fh

SYMBOL_STACK_INDEX: 3

SYMBOL_NAME: nt!MiIdentifyPfn+26f

FOLLOWUP_NAME: MachineOwner

MODULE_NAME: nt

DEBUG_FLR_IMAGE_TIMESTAMP: 4e02aaa3

IMAGE_NAME: memory_corruption

FAILURE_BUCKET_ID: X64_0xA_nt!MiIdentifyPfn+26f

BUCKET_ID: X64_0xA_nt!MiIdentifyPfn+26f

Followup: MachineOwner

 

[EVGA_MatthewH]

IRQ NOT EQUAL OR LESS is usually memory related. Do you have the memory timings, voltage, and speed manually set in the BIOS? If so try memtest (memtest.org) to ensure its working properly.

 

[fixbsod]

I have run like 8 full passes of the RAM w/MemTest and not found a single error, I have not yet run a stick at a time but I have run it with the RAM in different slots (rearranging the sticks).

The RAM did not want to work at 1866 initially despite all the correct settings and timings in the BIOS -- was causing occasional hardlocks and repeatable hardlocks w/Prime 95 during blend test. I slowed the RAM to 1600 w/XMP profile using the same stock 9-10-9-27-2T timings w/1.5V which eliminated all hardlocks since. Also the RAM is set to standard performance enhance, basically disabled, as the other options are turbo or extreme.

However the system has been stable under extreme load at 1600 -- Prime 95 12+ straight hours, IBT max (all RAM) 12+ straight hours, F@H 12+ straight hours, 120+hrs total gaming, and still not finding 1 error in MemTest with like 16 total hours (4 hours in a row then another solid 12 hours another day after rearranging the RAM). So I am dubious about it being a RAM issue. Esp, as all but 1 BSOD were with the computer idling doing basically nothing (torrenting). Even the 1 BSOD in game -- BioShock 2 was with the game paused, and it is my understanding that as it is a 32-bit game it cannot utilize more than 4GB of RAM.

The only other thought I have besides it being some driver issue or the RAM itself is that it could be the RAM voltage ?? I noticed that while VCCIO voltage (QPI/Vtt for the Gigabyte board) for many SB setups are in the 1.0x - 1.1x range that the read profile value is 1.20v and that is the current entered value. I did try 1.08v for a few hours and while this is supposedly only the RAM controller voltage, CPU-Z reported a MUCH lower voltage of 1.0x vs 1.2 for vcore -- I did not touch vcore. I did seem to get the same temps in P95 tho, so am thinking CPU-Z was reading the wrong value as I would have expected lower temps if it was really that much lower vcore. I am thinking about actually setting this to auto as perhaps there is too much voltage during minimal activity which is why it seems to happen when idling but not during use.

I have also heard that 1.5v may be insufficient for 1600+ RAM, especially when all 4 slots are used which could have been the reason for the hardlocks at 1866. So maybe both lowering VCCIO (1.08 or AUTO) and also slightly bumping VDIMM (1.55v?)

I could try lowering the RAM again to 1333...

I'm tempted to rip open the RAM I bought off of Newegg -- G.Skill Ripjaws DDR3-1600 9-9-9-24 1T 1.5V, and 'edit' -- they are XMP according to the modules, tho not detailed as such on newegg's webpage. Next BSOD they are likely going in.

IRQ NOT EQUAL OR LESS is usually memory related. Do you have the memory timings, voltage, and speed manually set in the BIOS? If so try memtest (memtest.org) to ensure its working properly.

 

[denis280]

try with one stick of ram first.

 

[Puppies04]

Do yourself a favour, pull 2 of the ram DIMMS out and set the remaining 8GB to 1600mhz. Post back if the problem still occurs.

 

[fixbsod]

18 hours+ no issues running Intel Burn Test using maximum RAM (16GB) -- something like 80 passes before I stopped it. Again, I have not been able to make the computer BSOD using IBT, P95, folding@Home, or any other stress tester/burn in tool.

I did some more tinkering as I noticed the issue seemed to happen weekly, thinking it could be some kind of auto-update/checker or even the weekly cable modem lease renewal that could be doing it. I disabled some auto-updating checks -- noticed Java 32-bit has a weekly update, and also disabled the evga Precision updater. I did also find some fairly outdated Marvell SATA drivers for the onboard SATA as well as outdated Realtek LAN adapter drivers and so updated both of those without issue.

Still no BSOD since the one on 1/24/12...tho I guess I should be expecting one like 2/1 or so...