Windows 8 BSOD, SPECIAL_POOL_DETECTED_MEMORY_CORRUPTION

becpc

New Member
Joined
Jul 27, 2015
Hi,

Recently, I got a new computer but I have been getting the blue screen of death (BSOD) almost since I got it. I thought that maybe it was related to a driver of one of the PCI cards connected to the computer but recently I removed all non-essential components (4 PCI cards and two extra HDD), and it was still happening. I ran memtest86 to test the RAM sticks and they gave some errors so I replaced them, but I'm still getting BSOD although not as frequent as before (new ones didn't show any errors on Memtest after 4 runs). At this point, I'm not sure what else to test and I have attempted to figure out the issues through the generated dump files but without success.

Note:Only the last dump files (since 07/24/15) correspond to the current configuration, the previous ones were generated when the previous faulty RAM sticks were used.

PC specs:

GIGABYTE GA-P85-D3 LGA 1150
Intel i7 4790
GSKill Ripjaws X 16 GB
Samsung 840 Evo 120 GB SSD
Corsair CX430M Power supply

OS

Windows 8.1 Pro (64x)

Attempted solutions

OS installation from scratch (x2)
Windows has lastest updates
All drivers are up to date

Tests (in order):

Windows memory test -> no errors detected
Memtest86 -> no errors on current RAM sticks
sfc /scannow -> no issues were detected
Ran Prime85 -> not immediately but computer gave BSOD
Stress computer by running several programs, and playing many videos -> not immediately but computer gave BSOD

Thank you,
BEC
 

Attachments

  • windowsforum.zip
    1.6 MB · Views: 605
Code:
*******************************************************************************
*                                                                             *
*                        Bugcheck Analysis                                    *
*                                                                             *
*******************************************************************************

Use !analyze -v to get detailed debugging information.

BugCheck C1, {ffffcf8256df0f80, ffffcf8256df0750, c74078, 32}

Probably caused by : memory_corruption ( nt!MiCheckSpecialPoolSlop+8a )

Followup: MachineOwner
Hi,
Bugcheck C1 is usually a driver related issue although I can see quite a few NTFS errors in the Stack Trace. As we have a number of possible suspects it's prudent to run a few tests. Also if you have the driver verifier running please disable it

Please run a chkdsk. This guide will show you how although whichever method you use please ensure you scan for and repair broken sectors:
The How-To Geek Guide to Using Check Disk in Windows 7, 8, or 10

As your machine hasn't been very stable running Prime95 just check all is well with your CPU with Intel's diagnostic tool:
Intel® Download Center

I found this driver running in the driver stack:
AgPciMem.sys Tue Jan 13 23:44:25 2015: 64-bit Agilent PCI Memory Allocation Driver. Relating to the Bugcheck above, although I saw no sign of this driver in the stack text i would be extremely suspicious of it simply because of what it does. If the bsod continues it might be worth uninstalling this and see if it improves things.

niede.sys Tue Jun 15 23:15:38 2010: National Instruments Corporation Ethernet Device Enumerator Driver. 2010 is a tad old for windows 8/8.1 please try and update if possible:
Support - National Instruments

I see you have the Intel extreme tuning utility installed. If your using it to overclock or stree the machine in anyway please stop and return to default values until we discover the source of the bsod.

BiosVersion = FA
BiosReleaseDate = 02/27/2014
SystemManufacturer = Gigabyte Technology Co., Ltd.
SystemProductName = P85-D3

Your Bios is slightly out of date and the next version includes stability updates:
http://www.gigabyte.com/products/product-page.aspx?pid=4977#bios

Please post any new dump files.
 
Hi kemical,

Thank you for your reply. Attached are new file dumps. I'll address your points in order and then add some observations:

1. I ran chkdsk on windows, no errors. Also I ran it from the console, when the computer restarted ran the test and spend some time on 13% but it finished and didn't raise any errors.

2. Intel's diagnostic tool ran fine, and every test was passed.

3. The driver (AgPciMem.sys) is for controlling one of the PCI cards that I normally use in this computer (at the moment, all PCI cards are out), and it supposed to be up-to-date for Windows 8. This PCI card is one of the reasons I have this computer, so I'm hesitant about uninstalling the driver but I'll do it if things don't improve. Also, I searched for a newer National Instrument driver but couldn't find one, I'll keep searching.

4. I don't have any overclock utility on Windows, but on BIOS there are there's a Intel turbo boost option. It was set to auto by default, I disabled it.

5. Since it seems like updating the BIOS can be a bit "touchy", I'll delay the update for now.

Observations and test:
Just to be clear, I'll talk about RAM position by that I mean there are four slots on my motherboard: Channel A DDR3_2 (Position 2), DDR3_4 (Position 4); and Channel B DDR3_1 (Position 1), DDR3_3 (Position 3)
  • BSOD three times without any apparent reason.
  • Ran Prime95, BSOD after 5-10 min (Two RAM sticks, RAM Positions 1&3).
  • Removed one RAM stick, left one on Position 1. Ran Prime95 for two hours, gave an error on one CPU (no. 5) but no BSOD. Stop scan manually.
  • Swapped RAM sticks, same position (Pos 1). Ran Prime95 for two hours, no errors. Stop scan manually.
  • Intel Diagnostic Tool x64 passed all tests (One RAM stick, Pos 1)
  • Chkdsk on windows, no errors (I had ran it before and haven’t seen any errors)
  • Chkdsk on terminal, spent some time on 13% but it finished and didn’t say anything
  • Prime95, 20 min into the test gave errors in different CPUs (4 different CPUs) (Two RAM sticks used, Positions 1&3)
hardware failure detected
rounding was a 0.5, expected less than 0.4
No BSOD, I stopped it to do other tests​
  • Prime95, 0 min (just at the start!) into the test gave errors in all CPUs (Two RAM sticks used, Positions 1&2)
hardware failure detected
rounding was a 0.5, expected less than 0.4
No BSOD, I stopped it to do other tests​
  • Prime95, few minutes (~10 min) no errors (One RAM stick, Position 2). I stopped it to do other tests.
  • Prime95, 30 min no errors, left to run overnight (Two RAM sticks, Positions 1&4)
I'm sorry for not letting run some of the tests for a longer time, but I only have access to the computer for some hours during the day. Some of the results give me the feeling that there may be something wrong with the RAM slots/motherboard but it's not clear and not sure how to test for it.

BEC
 

Attachments

  • new_filedumps.zip
    119.2 KB · Views: 292
Code:
*******************************************************************************
*                                                                             *
*                        Bugcheck Analysis                                    *
*                                                                             *
*******************************************************************************

Use !analyze -v to get detailed debugging information.

BugCheck C1, {ffffcf806922ef80, ffffcf806922ea90, abc078, 32}

Probably caused by : dxgmms1.sys ( dxgmms1!VIDMM_GLOBAL::CloseLocalAllocation+2b1 )

Followup: MachineOwner
Hi,
same again although this time I can see that graphics subsystem was also mentioned. Now depending on what you wrote back to me my next suggestion was going to be the RAM slots or RAM. Judging from what you've written here:
Observations and test:
Just to be clear, I'll talk about RAM position by that I mean there are four slots on my motherboard: Channel A DDR3_2 (Position 2), DDR3_4 (Position 4); and Channel B DDR3_1 (Position 1), DDR3_3 (Position 3)
  • BSOD three times without any apparent reason.
  • Ran Prime95, BSOD after 5-10 min (Two RAM sticks, RAM Positions 1&3).
  • Removed one RAM stick, left one on Position 1. Ran Prime95 for two hours, gave an error on one CPU (no. 5) but no BSOD. Stop scan manually.
  • Swapped RAM sticks, same position (Pos 1). Ran Prime95 for two hours, no errors. Stop scan manually.
  • Intel Diagnostic Tool x64 passed all tests (One RAM stick, Pos 1)
  • Chkdsk on windows, no errors (I had ran it before and haven’t seen any errors)
  • Chkdsk on terminal, spent some time on 13% but it finished and didn’t say anything
  • Prime95, 20 min into the test gave errors in different CPUs (4 different CPUs) (Two RAM sticks used, Positions 1&3)
It looks to me that either a stick or the slot is still faulty. I know you tested these sticks with Memtest86 but sometimes I've known Memtest86 not to pick anything up (in the time it was run meaning if it was ran for longer it's possible that a fault may have been found) but sticks have still been faulty. This is discovered by testing like you did above. If I read it correctly you got CPU errors using one stick and not the other.
Also what are your temperatures like when running Prime95?

If you can try running on the stick which gave no errors and see if the bsod's return. If they don't then you know either the other stick is faulty or the slot it was used in is..
 
Hi kemical,

I ran Memtest86 again to test the RAM sticks, and they gave errors. I found that hard to believe since these sticks are new, and I just got them since my previous RAM stick showed errors a few days ago. I decided to update the BIOS, and it *seems* like that was the problem, no BSOD so far. I have ran Memtest85, no memory errors, and Prime95, again no errors in a couple of hours.

Also I noticed CPU Vcore is lower now, before it went up to a max of 1.16V, and now it only goes to ~1 V, maybe that was causing instability.

Hopefully I don't have to come back with the same problem :)

Thank you,
BEC
 
Hi BEC,
I really appreciate you updating your thread and am glad you have now found the issue. We never close threads so if you ever have a similar issue or anything we can help with for that matter just pop back.. :)
 
Back
Top Bottom