PDA

View Full Version : GeForce 6800 GT causes Spontaneous CPU Reboots.


jcamus
10-12-04, 03:28 AM
I am having terrible problems with my brand new system built entirely around the nVidia GeForce 6800 GT. I have intermittent spontaneous reboots when playing 3D games.

Custom Built CompUSA System:
* AMD Athlon 3200+ (2200 MHz) with 200MHz external bus
* BIOS: Award Software International, Inc. FC 03/01/2004
* 1024 MB physical RAM (Kingston 400 speed RAM)
* GigaByte K7 Triton series nForce 2 motherboard
* Windows XP SP2 setup on 08/28/2004 (legal, registered copy)
* 500 Watt ATX ULTRA Power Supply

What I Have Done So Far:
Of course I installed the most current nForce2 drivers and Detonator drivers. Tried some "leaked" newer drivers off Guru3d. Finally uninstalled all drivers, RMA'd the 6800 GT and installed a brand new one and reinstalled the official drivers.

Took system in to CompUSA where they said "it is your power supply" and upgraded from 400 watt power supply to 500 watt power supply. Also upgraded CPU heat sink to copper; CPU and GPU running real cool. CPU = 56 degrees and GPU = 58 degrees as I write this.

After all this, I play "Evil Genius" and my system spontaneously reboots anywhere from 7 to 20 minutes after I start playing. Graphics freeze for 10 - 20 seconds before blackout and reboot. Windows unhelpfully offers to report a general, unsolved error with vague references to nVidia and graphics driver.

Anybody have some idea what might be so horribly wrong with my system build? This is driving me NUTS! And no, I haven't overclocked anything.

--jcamus

|MaguS|
10-12-04, 03:31 AM
First of all, Did you disable DEP in SP2? (Sp2 sucks for gaming). Also wattage means jack when it comes to PSU, if the quality isn't there it could be 1k watts and still put out less then a 400w PSU.

Have you tried turning Fastwrite on and off to see if that does something? How about Bios update on your Motherboard?

Reliant
10-12-04, 03:33 AM
1. Your CPU is overheating due to bad contact with HSF and or bad case cooling. 56C is actually pretty high Idle temp you might have put the thermal paste on too thick.

2. Your Generic power supply isn't up to snuff on the 12 volt rail.

Kickus_assius
10-12-04, 03:35 AM
Faulty RAM maybe a problem as well.

racingx
10-12-04, 06:53 AM
I would think it is faulty RAM that is causing this also. If you have 2 sticks of RAM installed try taking out one stick and see if your problem clears up. Also go into your mobo bios and try relaxing the memory timings some also. A couple of other things you could try is making sure AGP Fastwrites is disabled, or maybe bumping up the AGP voltage in the bios.

mattkenn_4545
10-12-04, 06:08 PM
I agree that your temps for an athlon xp should be no higher that about 50c. Look at my sig but then again I cool the thing with ac but before if I got any higher that about 45 I would get unstable.

EciDemon
10-12-04, 06:21 PM
My vote goes out to ram as i experienced simalar problems when I upgraded.
+pov 6800 +new hdd +512mb ram

Problems I got was lots of BSOD and drop to desktop, games would play for less then five minutes, some times more.

I really thought it was the Gfx card so I returned it and those guys in the shop wanted to test it first.

When I got home and installed my old fx5200u i noticed that the problems were still there. Thats when I started to really check the error logs and note down the error codes, then I did a lot of googleing and came to the conclution that it might be the ram, I individually thested the new sting and the old stick. When I did that they worked perfectly. They just were not compatible with echother.
So I picked up my gfx card and another 512mb ram so they are of the same brand, Now things work perfectly again.

TacT
10-12-04, 07:03 PM
I agree that your temps for an athlon xp should be no higher that about 50c. Look at my sig but then again I cool the thing with ac but before if I got any higher that about 45 I would get unstable.

an xp chip should be fine and stable until about the 55-60C range I would have thought, I know mine is.

Your whole system is overheating and causing lockups I'd hazard a guess. 56C xp chip and a 58C gpu chip is not cool at all for idle temps. Take the side of the case off and stick a big fan or hairdryer on full blast blowing cold air over your components and try running a game again. Does it pass a prime95 torture test? It could be faulty ram too considering it's randomly rebooting, have you tried running memtest?
Have you tried running one stick of ram at a time to determine if one might be faulty? Are you sure the cpu hsf is well fitted with just enough AS5 or alternative thermal compound with good contact for heat dissipation? Does your motherboard need a bios update perhaps? Have you tried asking yourself any of these questions?

Last of all, what in gods name is this 550W ULTRA SUPAR PSU you have purchased? Is it made by nintendo? If it is it could also be the problem ...

photophreak314
10-12-04, 07:55 PM
CHeck your ram. I had the exact same problem, and it turned out that my bios was forcing a cas timing of 3 instead of 2.5 @ 200mhz. (Samsung DDR400 Ram, pretty cheap, but really good once I got it figured out)

Download and run Prime95, and run the torture test. I used to fail in less than a minute, but now can run it indefinately.

I figured out the RAM (I didn't know anything about timings before hand) by running it at 133mhz, which worked, and then 166, which worked, and I got up to 178mhz, which is when the cas timing would change to 3, so I set it to 200, forced the timing to 2.5, and all is good.

jcamus
10-12-04, 09:44 PM
EVERYBODY: Thanks for replies so far; I am thinking the RAM is the problem right now, but don't know yet. I think those of you guys harping on temperature are off base; this system has 6 or 7 fans going including top-case, side-case, rear-case, dual power fans and of course onboard CPU/GPU fans. There is PLENTY of airflow and the CPU isn't that hot.

First of all, Did you disable DEP in SP2? (Sp2 sucks for gaming). Also wattage means jack when it comes to PSU, if the quality isn't there it could be 1k watts and still put out less then a 400w PSU.

Have you tried turning Fastwrite on and off to see if that does something?

I've already messed with Fastwrites; no help at all. DEP says it is "only enabled for essential Windows services" already so I'm not sure how that applies. I can set it to opt-out for individual applications but I have no idea how to just "disable" DEP for the system.

photophreak314: I will download Prime95 and try the stress test maybe all night tonight. I don't know much about BIOS settings but this is 400MHz speed I believe and I guess I could slow it down and try that???

Maybe I will try the 1-RAM stick thing too.

Thanks and I'll try to keep bashing this crappy build into working. :bash:

--jcamus

Zetto
10-12-04, 09:46 PM
Try setting up a desktop fan to blow right into ur case with side panel off. See if it helps. It does for me :D But then, I oc'ing the hell out of CPU and memory ;)

Reliant
10-12-04, 10:15 PM
56C Idle is not a good CPU temp especially with as many fans as you have. 70C is the point at which the CPU reboots the system.

jcamus
10-13-04, 12:21 AM
Would someone please explain this output from Prime95 to me?

It says all these "Self-test passed" for 1024K, 8K, 10K, 896K, 768K, 12K, 14K, and then it has a paragraph saying this:

Test 1, 6500 Lucas-Lehmer iterations of M12451841 using 640K FFT length.
FATAL ERROR: Rounding was 0.449835968, expected less than 0.4
Hardware failure detected, consult stress.txt file.
Torture test ran 2 hours, 1 minutes -- 1 errors, 0 warnings.
Execution halted.

What does this mean for my Kingston 400 ram? The computer didn't crash, just the test stopped.

--jcamus

racingx
10-13-04, 07:19 AM
Have you tried testing one stick at a time yet? Also sometimes rasing the VDIMM can help acheive stability for your RAM. Just dont overdue it because you will fry your RAM.