Solved BSOD when playing BF3 (Battlefield 3), error 0x000000fc

[James7679]

Yes, please run the JCGriff report again.

 

[rikbrouwer]

Ran the JCGriff report tool, the output is attached to this post.

 

[James7679]

Hmm, still looks like the GPU drivers are crashing. Please stress the Video Memory first, then the Video Card itself. Please monitor Voltages with HWMonitor and temps with CoreTemp. Post back the results of both tests along with voltages and temps during testing.

 

[rikbrouwer]

Hi James. For some reason the video memory test won't run. I get this log:

WARNING: Your card can be not compatible with this test! If multiple errors will found, try test with onscreen buffer usage or wthout of ignoring colors bit mask.
Changing video mode to 640x480x16...OK
WARNING: Do not place other windows over program window!
INFORMATION: Only one mode can be tested now, all others will not match.
[24-6-2012 10:11:00] Test started for "Primair beeldschermstuurprogramma (AMD Radeon HD 6900 Series)"...
Trying 16bpp RGB:565 mode...OK
PATTERN NOT FOUND
Trying 16bpp RGB:555 mode...NOT MATCH WITH CURRENT
Trying 16bpp BGR:565 mode...NOT MATCH WITH CURRENT
Trying 32bpp RGB:888 mode...NOT MATCH WITH CURRENT
Trying 32bpp BGR:888 mode...NOT MATCH WITH CURRENT
NO PASS COMPLETED

I tried to switch the "onscreen buffer usage" and "ignoring colors bit mask", but this didn't make a difference. Oh, and I had to start vmt.exe with parameter "2048" because this program thought the 2GB of video memory had to be wrong. Guess it's a bit outdated..

Anyhow, I've successfully ran the Furmark test, however had problems with HWMonitor. It immediately stopped running while initializing the program. Therefore I've used Speedfan to monitor the voltage. Core Temp required me to install Ask.com crap, so I refused to install that one. Since speedfan also monitors temps, I figured that would do the job as well.

Attached is the GPU temp and voltage, as well as the Furmark screenshot, right before I stopped the test. No problems here.

min temp: 57C (135F)
max temp: 90C (194F)
min vddc: 1.000V
max vddc: 1.100V

 

[James7679]

Please upload the screens for CPUz.

Also, please try underclocking the GPU and see if you can find stability. I see you unlocked your 6950 to a 6970. Were you having these problems before you flashed the BIOS? I have found that doing the flash can cause some instability. You might try reverting the GPU back to it's stock values. Doing so, is up to you. Doing any BIOS flash is a risky one and can result in damaged or inoperable components.

 

[rikbrouwer]

Please upload the screens for CPUz.

I've attached a report sheet, is that what you meant?

I see you unlocked your 6950 to a 6970. Were you having these problems before you flashed the BIOS?

Yes, the GPU has the 6970 BIOS flash. However the card has a small switch, which enables you to switch between two BIOS's, in case you **** up your flash. I'm using the stock BIOS again, since your first reply.

Before I even had this card in my computer, I had a HD5770. This one gave annoying freezes up to 30 seconds while playing MW2 and MW3 (often resulting in a disconnect/timeout). The HD6950 is actually my little brothers GPU, which he flashed and used without problems for over a year. He then sold his entire computer to my parent, in order to buy a notebook. Lucky me
So I've switched my HD5770 with this HD6950 (in 6970 mode), but the problems increased, resulting in the BSOD's.

Next I'll try to underclock the GPU then..

edit: added CPU-Z screenshot. Easier than the text file

 

[James7679]

Okay, thanks, I also need the memory, speed and mainboard as well. Also, go ahead and upload the graphics tab as well.

We are gonna stress your system and check some things.

Please go into your BIOS ans write down the values for your Voltages, they should be under PC health status. 3.3v= 5v= 12v=

Download and install;
Memtest86+ - Advanced Memory Diagnostic Tool Tutorial http://www.sevenforums.com/tutorials/105647-ram-test-memtest86.html
CPUID - System & hardware benchmark, monitoring, reporting
Core Temp Be sure to use the free version, not Pro.
Free Software - GIMPS Tutorial Torture test your CPU with Prime95

First, run Memtest86, let it run overnight. If you get any errors, remove all but one stick and test each stick, in each slot, independently. Take note of the behavior. If you get an error, move to the next slot/stick. Again, take notes.

Next, stress test the CPU using Prime95. Take a screen shot of the CoreTemp window and HWMonitor and upload it.

Also, BF3 is what causes the crash generally, right? Did the GPU have the same problem when it was in your brother's PC? If not, I would say that there is obviously some other factors involved here. Play BF3 for a bit and take a screen shot of CoreTemp and HWMonitor and upload those as well.

 

[rikbrouwer]

Okay, thanks, I also need the memory, speed and mainboard as well. Also, go ahead and upload the graphics tab as well.

Attached them to this post

Please go into your BIOS ans write down the values for your Voltages, they should be under PC health status. 3.3v= 5v= 12v=

(Reading in Speedfan | Reading in BIOS)
3.3v = 3.31v | 3.312v
5v = 5.02v | 5.019v
12v = 11.88v | 11.93v

We are gonna stress your system and check some things.

Download and install;
{..}

First, run Memtest86, let it run overnight. If you get any errors, remove all but one stick and test each stick, in each slot, independently. Take note of the behavior. If you get an error, move to the next slot/stick. Again, take notes.

Next, stress test the CPU using Prime95. Take a screen shot of the CoreTemp window and HWMonitor and upload it.

I've already ran Memtest for 9 hours right after opening this topic (see screenshot). Should I run it again?

Also, already did extensive Prime95 testing, because before I asked for help here, I tried changing my CPU fan (stock to Scythe Mugen II rev.B), and I've added some case fans. See screenshot from Excel table for the different temp readings from CPU and MB under stress as well as idle. The left side of the table shows different case fan configurations. The first results (line 3) is the stock configuration.

Also, BF3 is what causes the crash generally, right? Did the GPU have the same problem when it was in your brother's PC? If not, I would say that there is obviously some other factors involved here. Play BF3 for a bit and take a screen shot of CoreTemp and HWMonitor and upload those as well.

It happens very occasionally outside of BF3, but yes, overall it seems BF3 triggers the problem, or BF3 pushes my hardware to its boundaries. However I can't reproduce the error by stressing. I've already monitored the temps while playing BF3 for many hours and on many different days. GPU gets hot, around 80C (176F). CPU stays cool, not exceeding 40C (104F). Problem is that when my computer suddenly gives an BSOD, I can't see the temp readings. But they do not fluctuate much, so I don't think it directly causes the problems. I'll upload another screenshot when I'm playing again. This week my finals, so too busy studying to play BF3.

(By the way; Can I use Speedfan for the temp readings? Or do CoreTemp and HWMonitor have something special?) Nevermind, already had both as a portable app on my Dropbox.

Oh, and my brother never played BF3 or one of its predecessors, so we don't know if a similar problem would have been present on his system.

 

[rikbrouwer]

Did another Prime95 Torture test for 35 minutes. See attachment
(cpu fan RPM readings are wrong.. obviously)

 

[James7679]

Let me know how things progress after rolling back the BIOS and underclocking the GPU goes. There are a couple of more things that you can try, but it's better to allow each step to confirm or deny success. All of your voltages and temps look great. Stability in P95 and Furmark are good. But BF3 pushes systems far beyond what normal stress testing will do. After I check my overclocks with P95 and Kombustor, I always try BF3 for a bit. I have had weaknesses that BF3 found, which the other programs didn't.

 

[rikbrouwer]

Problem is solved:

In the summer I've bought a new SSD (Samsung Series 830 - 128GB) because the 60GB I had became too small, and a new motherboard, because I wanted AHCI and sata600. Besides, I was hoping these replacements would solve the problem. Well.. they didn't.

Two weeks ago I found a nice deal for 16GB of RAM, which I bought.. And this seems to have solved my problem of endless blue screens and crashes. Strange, because I had tested the RAM for errors.

Anyhow, thank you all for your help. Happy that I can trust my machine again!