8800GTS 640 BSOD/random reboots

supraw

Distinguished
Jan 17, 2008
4
0
18,510
I am getting random restarts during games. I can get about 1 hour of gameplay in UT3. If I'm not running any 3D apps or games, my system seems stable, which is why I'm posting this here.

Usually it just reboots - I don't see a BSOD.. but last time this happened it BSOD'd and I got a message upon restart (more on that later)

Here are my specs:

EVGA 8800 GTS 640
EVGA nForce590 AMD SLI AM2 board
AMD X2 64bit 4600+
2x OCZ 1025MB PC6400 DDR2 800MHz dual channel RAM
PC Power & Cooling Silencer 750W
2x Seagate 320GB SATA 7200 running in RAID-1

Version Numbers:

EVGA display driver: 169.28 (6.14.11.6928)
GPU BIOS: 60.80.0A.00.01
mobo BIOS: not sure how to get this atm, but I juuust flashed it with the latest BIOS update I downloaded from EVGA for my board. How do I find out what version I have and what version should I have?
nForce IDE: 6.87 (5.10.2600.0687)
nForce networking: 60.15 (1.00.01.06015)
UT3 - patched (there's only 1 patch out right now)
CPU - 1.3.2 drivers installed as well as the dual core optimizer

I have stress-tested my RAM both with mem86 and prime95 and am confident that my RAM is good.

The error codes I received upon restart from the BSOD are as follows:

BCCode : 1000008e BCP1 : 80000003 BCP2 : 8051FFFD BCP3 : A7313C98
BCP4 : 00000000 OSVer : 5_1_2600 SP : 2_0 Product : 256_1

I have used Windbg to try to make sense of the dump file, but it did not tell me much. Here are some experts.. it didn't help me, but some of you might make sense of this:

EXCEPTION_CODE: (HRESULT) 0x80000003 (2147483651) - One or more arguments are invalid

FAULTING_IP:
nt!MiDereferenceSegmentThread+a2
8051fffd cc int 3

TRAP_FRAME: a7313c98 -- (.trap 0xffffffffa7313c98)

DEFAULT_BUCKET_ID: DRIVER_FAULT

BUGCHECK_STR: 0x8E

PROCESS_NAME: UT3.exe

LAST_CONTROL_TRANSFER: from 80543908 to 8051fffe

STACK_TEXT:
a7313d4c 80543908 00000001 fc8eee94 00000001 nt!MiDereferenceSegmentThread+0xa0
a7313d64 7c90d800 badb0d00 7c90eb94 00000000 nt!`string'+0x18
WARNING: Frame IP not in any known module. Following frames may be wrong.
Invalid unloaded module list count

IMAGE_NAME: memory_corruption

I have also tried to use verifier (comes with XP) to test my drivers, and they all checked out. The Windbg results seem to indicate that this is a driver problem. Is this really that simple?

I have googled the error codes I got, and I ended up reading a LOT about memory problems. However, my memory checks out, like I said before.

It seems to me like Unreal Tournament is trying to reference an invalid memory block, causing the crash. Is this simply a UT problem? I have scoured the UT forums but have been unable to find anything there, and I have patched the game.. which seemed to help, but then .. boom.. crash.

I'm not sure if this will help, but I figure, the more information I include, the better. Here are my CPU/GPU temps. Load is low right now and has been for a while. I was running stress tests earlier, but have been just browsing the web for the past 20 minutes or so:

CPU: 102F / 39C
GPU: 140F / 60C

If you see any issues with any of the numbers I have provided, but don't think you can really help.. by any means reply and let me know if something doesn't look right!

I was also reading up on my CPU and have seen some people installing a hack/patch. Apparently there are some issues with AMD dual-core chips. AMD used to have a fix on their site, which many threads pointed to, but it has since been removed. I assumed that their latest drivers fixed this problem directly, so I did not apply any additional hacks/patches.

I just installed XP and the whole shebang on monday, and that's when the problems started. Actually, I also had them before that, but they were slightly different. I suspected my PS and got a new one. My system seems way more stable, ie. it doesn't freeze during non-3d activities.

Based on what I have read it could be a mobo problem but tons of people have had similar problems as me, and they often don't turn out to be a hardware problem. Exactly the same tyep of problem was often fixed by doing seemingly unrelated things like changing the paging size, re-seating your PCI cards into different slots, upping memory voltage to 2.1V (Mine's at 1.9V), etc. Seems like something's not playing nice with something else, and I want to figure out what it is.

I should also mention that nTune freezes for me from time to time, usually on the Mobo/GPU screen with the temps.

Here's what I've done since:

1. Flashed mobo BIOS
2. Restarted
3. Loaded default values in BIOS (as per instructions)
4. Removed display drivers
5. Booted into safe mode, used drive sweeper
6. Boot into not-so-safe mode, installed 169.28 display drivers from EVGA.COM
7. Installed Crysis
8. Game caused machine to reboot in INTRO (it was not pre-rendered, it was live 3D)

I keep getting UT3 blue screens and reboots (sometimes the blue screen will show up and reboot right away, sometimes I will not see any blue and it will reboot. Keep in mind that I have it set up not to reboot on BSODs)

A couple times, a couple minutes before the reboot (and this happened twice), my computer would "hiccup" during UT3, but not crash. It'd slow down, speed up again, slow down, speed up, 4-8 times, all that in around a second or two.. producing some sort of "chuggy" sound. It was sudden and quick so I didn't have the foresight to see if that was just in-game sound (it seemed like it was that), or if there were accompanying sounds coming from my case. After that happened, I was able to continue my game, no reboot... but then 10-20 minutes later I'd get a similar thing, just a bit more intense, and it would actually reboot.

One of the blue screens actually stayed there so I could write down this:

IRQL_NOT_LESS_OR_EQUAL

STOP: 0x0000000A (0x0A330005, 0x00000002, 0x00000001, 0x804FEEC0)


Upon some googling I've discovered that this could be due to a driver conflict between two devices. I have not yet looked into this but here is what I'm going to try next:

1. Remove all motherboard drivers (including onboard sound and onboard ethernet)
2. Safemode - drive sweeper
3. Try playing UT3 without installing any drivers
4. Install base mobo drivers, no network or sound
5. Try to play UT3 for 3 hours straight (most i've been able to do is an hour and a half)

Any help at all would be appreciated.

I should note that I couldn't open the last memory dump file because it was corrupted. During most reboots, one doesn't even get generated.
 

krillz

Distinguished
Jan 2, 2008
53
0
18,630
My problem with a 640 XFX GTS card, was caused by my mobo and built in sound card, turned that off in the bios and added a new sound card to my box. The wierd memory BSOD now vanished, instead I started to get random IRQL_NOT_EQUAL_OR_LESS crashes randomly.

So I just started shutting down things in the bios that I didn't use, freeing up IRQLS. Still occasionally I got strange BSOD reffering back to the memory.
Restarted and ran some tests. Then it suddenly stoped. No craches in 2-3 days.

Then I made a new fresh OS install, this time I downloaded all the latest drivers for everything, intel chipset drivers etc from intels page. As little as I could from the mobos site and graphic card's site.

Finally I got the system stable. Don't know if this will work or help you a bit to solving it. But it feels as I were experiencing the same symptons as you do now.

Also I flashed the graphic card's bios to get the fan to adjust, and one suggested to pull the fan speed up with rivatuner or some similar application to prevent it from overheating.

And I also changed my PSU as it was giving to little power causing the video card to shut down from time to time while gaming.
 

supraw

Distinguished
Jan 17, 2008
4
0
18,510
Well, I did do something in my BIOS..

and as a result was able to play Crysis on high settings for 1 hour with no slowdowns and no reboots.. no BSODs.. at all

What did I do?

Upped my RAM voltage from 1.9 to 2.1.. Now why the hell did that work?
 

supraw

Distinguished
Jan 17, 2008
4
0
18,510
Oh, I also "stress tested" by playing UT3 for 4 hours, high settings, no slowdowns, no artifacts, no reboots
 

supraw

Distinguished
Jan 17, 2008
4
0
18,510
That is useful feedback, thanks.

I am still not sure why my system worked fine when I had it set to 1.9V fo r 6 months, though.

Maybe the reboots didn't happen nearly as frequently because the games weren't as demanding and I often just play Civ.. which isn't as 3d intensive.. so maybe I just never noticed it.

But still, I am curious..