Random BSODs with demanding applications

wnieuwstraten · 26 Jan 2012

Tested both sticks of RAM on both slots. Five different tests:
Stick 1 in Slot 1: 10 hour, no error
Stick 2 in Slot 1: 10 hour, no error
Stick 1 in Slot 2: 1 hour, no error
Stick 2 in Slot 2: 1 hour, no error
Stick 1 in Slot 1, Stick 2 in Slot 2: 1 hour, no error.

Upon closing the testing program, I did get a BSOD. Different message to this one, might be useful or irrelevant/coincidence. I will test an overnight test of the double RAM sticks, to make sure I didn't do wrong.

MemtestG80 without errors
VMT with one error: "Error at [2C9B6C84]: must be 0800, but found 8800 (bits: 1000000000000000)"
Artifact locator without errors
CPU Stab test without errors

usasma · 26 Jan 2012

Well the only error was the VMT. And that's video.
The BSOD is a STOP 0x3B and it blames DirectX (a video item)

I'd have to suspect the video card. Can you try another one?

But I'd also have to wonder why you had Prime95 errors - and they didn't show up after moving the sticks around.
It may have been dust buildup around the slots, but that's just a guess.

BSOD BUGCHECK SUMMARY

Code:


Loading Dump File [C:\Users\FUBAR\_jcgriff2_\dbug\__Kernel__\012612-11575-01.dmp]
Built by: 7601.17640.amd64fre.win7sp1_gdr.110622-1506
Debug session time: Thu Jan 26 15:15:40.008 2012 (UTC - 5:00)
System Uptime: 0 days 0:43:33.756
Probably caused by : dxgmms1.sys ( dxgmms1!VIDMM_GLOBAL::CloseOneAllocation+1e )
DEFAULT_BUCKET_ID:  VISTA_DRIVER_FAULT
BUGCHECK_STR:  0x3B
PROCESS_NAME:  firefox.exe
FAILURE_BUCKET_ID:  X64_0x3B_dxgmms1!VIDMM_GLOBAL::CloseOneAllocation+1e
Bugcheck code 0000003B
Arguments 00000000`c0000005 fffff880`101b13a6 fffff880`0b8240d0 00000000`00000000
BiosVersion = E7728MLN.207
BiosReleaseDate = 07/19/2011
CPUID:        "Intel(R) Core(TM) i7-2600 CPU @ 3.40GHz"
MaxSpeed:     3400
CurrentSpeed: 3392
¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨¨``

wnieuwstraten · 26 Jan 2012

If it had been building up dust, I was doing so at the manufacturer's. The Prime95 Blend test with both chips has been going at it for an hour now without any troubles. Good to know this probably isn't the problem. :)

I will do the only test from your link I haven't done yet: OCCT

Will update.

wnieuwstraten · 27 Jan 2012

OCCT test gave at high resolution zero errors (13 FPS avg, GPU 64 °C max).

Prime95 Blend test over 10 hours, no BSODs, did get three errors on Workers #1, #3 and #8 after respectively 2h25m, 2h22m, 2h32m. Test stopped with these cores but continued on the other cores.

I'll try running a game which gave me BSODs earlier, could it be that just unpluggin and replugging the RAM has helped the problem?

usasma · 27 Jan 2012

My first great embarrassment as a PC Tech involved someone else reseating the sticks after I'd given up.
Yes, it can make a difference! :0)

Prime 95 errors do indicate RAM - but running all 3 tests (and timing how long it takes to generate an error or a crash) is the ultimate test. I think that there's a hardware problem with one of the memory structures that Prime95 tests (CPU, memory controller, or RAM). I also have to wonder about a video card issue because of the error in the VMT.

wnieuwstraten · 28 Jan 2012

Hard to test now. I'm getting no more BSODs after hours of playing at high resolution. That is a good thing!

Thank you very much for your help!