nVIDIA GeForce 6800 GT Locking up my system

Message

ThElInK · Post by **ThElInK** » Sat Oct 29, 2005 7:14 am

I've recently upgraded to an XFX GeForce 6800 GT from my previous card, a GeForce FX 5200. Since having done so, I'm having problems playing games.

When starting Doom3 in "Ultra high" or "high" quality mode, the game locks up either seconds, or minutes, after the game begins. The whole system freezes, neither CTRL+ALT+BACKSPACE nor anything else works. Starting doom in "medium" quality mode seems to stop the problem from happening, although I haven't played the game for longer then 30mins. When starting Half Life 2, the game freezes a few seconds after the menu is displayed. With Counterstrike: Source, the game loads a map, then sends me back to the desktop.

Here's some information you may need:

Gentoo running amd64 on AMD Athlon 64 3000+ (Socket 939)
XFX GeForce 6800 GT (AGP)
1.5GB DDR RAM

cat /proc/driver/nvidia/agp/status:

Status: Enabled
Driver: AGPGART
AGP Rate: 8x
Fast Writes: Disabled
SBA: Enabled

Kernel: 2.6.12-gentoo-r6
nvidia-kernel: 1.0.7676
nvidia-glx: 1.0.7676-r1

I've also used nvidia-settings to check the card temperature. 15mins after boot, nvidia-settings reports:

Core Temperature: 44C
Ambient Temperature: 36C

I've searched through the forums for similar problems, and this has been reported one or two times, none of these people found a solution (or posted about it). I looked on my motherboard for capacitor decay (as described in another post) and haven't found any. I also don't have X Composite extensions enabled.

Any help would be greatly appreciated.

whig · Post by **whig** » Sat Oct 29, 2005 8:18 am

With a 6800 GT you will need a good powersupply, 400W at a guess. Check the voltages on your rails.

ThElInK · Post by **ThElInK** » Sat Oct 29, 2005 8:30 am

I have a 500W power supply - shouldn't be a problem

d13f00l · Post by **d13f00l** » Sat Oct 29, 2005 10:18 am

I just had this happen to me with an EVGA 6200, windows or linux as well.
I sent it back. A 5700 works fine in my PC, and that uses more power. Waiting on the replacement.

Uhh...I think that card has an extra power supply connector. You connected it, right?

ThElInK · Post by **ThElInK** » Sat Oct 29, 2005 3:31 pm

d13f00l wrote:Uhh...I think that card has an extra power supply connector. You connected it, right?

Yep, the card doesn't even work with the connector disconnected.

Since I posted, I've upgraded my nvidia-kernel and nvidia-glx, they're now both 1.0.7676-r1. I've also upgraded my kernel to 2.6.13-gentoo-r3, but the problem persists.

Aynjell · Post by **Aynjell** » Sat Oct 29, 2005 5:45 pm

Does it's built in bios speaker scream bloddy murder?

That is the scariest moment in setting up a PC, starting it up and hearing:

EEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEE!

bkzshabbaz · Post by **bkzshabbaz** » Sat Oct 29, 2005 6:13 pm

do you know anyone that has a 6800GT you can borrow? I have a 6800GT as well, and it hasnt given me any problems. While I was shopping for a 6800GT, I read alot of bad feedback on the XFX brand of cards.
I would try the following:
Isolate the problem by using another card, if the same problem appears, swap and use another PSU, if problem persists, run it on windows, if there is no problems with windows, you would know it was a software problem.

Thats at least what I would do.

Aynjell · Post by **Aynjell** » Sat Oct 29, 2005 7:48 pm

Okay, let's resolve this:

XFX, at least now, is high quality and is SUPPOSED to be in the top tier, with thier lifetime warranty. Thier cards shouldn't suck, and if they do, it would most likely be a bios problem (video bios), and flashing it with a bios from another vendor's card might resolve it. Bios flashing isn't as hard as some might think, but it's just as risky.

Moving on, what brand is your power supply? Your 500 watt unit may very well be giving you crappy power. Even if it's 500 watts of it. I'm not to clear on what it means, but apparently there's "Dirty" power that does no good for electronics. So, depending on what PSU you have, it may or may not be that.

Next up, what motherboard, and what chipset are you using? There are many reports that GeForce 6 and nForce 3 simply do not get along, ESPECIALLY when overclocked. Yes, I know it sounds wierd, but if you hadthe following components you'd notice that yoyur card would constantly stutter. Mine after a few changes from 2d to 3d rendering, and maybe when I change the frequency... It'll draw about 20 frames then lock for a a quick second. Rince and repeat...

Anyway, report back, maybe we can help.

ThElInK · Post by **ThElInK** » Sat Oct 29, 2005 8:40 pm

I've put a hard disk with Windows installed on it in this machine. I then installed Doom 3 and ran it in "Ultra High" mode. Although it was a bit jerky in places, the game ran fine without freezing (ultra high on Linux would freeze it after a few seconds). This seems to be a software problem.

Seondly, I agree with Aynjell. XFX cards are supposed to be top quality, why everyone keeps bashing them is beyond me. Although I think it's a software problem, here's the info you requested nontheless

Brand of PSU: 3XS Systems (Switching Power Supply)
Motherboard: Gigabyte GA-K8NS-939 with an nForce 3 Ultra chipset

I don't have any overclocking.

Again, I *highly* doubt this is a hardware problem: everything works fine from Windows.

Aynjell · Post by **Aynjell** » Sat Oct 29, 2005 9:23 pm

I want to stress two things here:

One, you are using an off brand power supply unit. See if you can pick up an OCZ Powerstrream, or modstream, or even some type of PC Power and Cooling unit. Both will most likely remove any shadow of a doubt that your PSU is failing you. Considering both Powerstream and PC Power and Cooling units were the first units to support SLI, not so much be approved by... but if I recall, anandtech was using powerstreams to test SLI setups, and with DFI boards, they are the best PSU to go with. This is verified by users, testing firms, DFI themselves... etc.

Next up, your XFX card is fine. So that is not a worry, but your nForce 3 chipset very well may be the culprit. Try turing off fast writes.

big_D · Post by **big_D** » Sun Oct 30, 2005 12:16 am

I don't think he has fast writes enabled. Turning off Side Band Addressing *may* help, although I've had no problems with either running a very similar set-up.

ThElInK · Post by **ThElInK** » Sun Oct 30, 2005 6:45 am

Do you really think a new PSU will help? I've noticed that one of these "high end" PSUs cost four times as much as a standard one - Locally, a normal PSU costs £20, whereas a Tagan High-End SLI/SATA/WHATEVER PSU costs £80. If my games weren't working in Windows, I'd go out and buy one. But, since my games are working in Windows, how likely is it this is actually a PSU problem?

I don't think he has fast writes enabled. Turning off Side Band Addressing *may* help, although I've had no problems with either running a very similar set-up.

Fast writes are not enabled. How do I disable Side Band Addressing?

Aynjell · Post by **Aynjell** » Sun Oct 30, 2005 3:43 pm

Fast Writes *SHOULD* be enabled by default.

ThElInK · Post by **ThElInK** » Sun Oct 30, 2005 4:19 pm

Aynjell wrote:Fast Writes *SHOULD* be enabled by default.

I enabled fast writes using the guide at http://gentoo-wiki.com/HARDWARE_Nvidia_ ... Addressing I *STILL* have the same results though - Doom 3 freezes a few seconds after gameplay beings at "ultra high" and "high" modes. HL2 freezes on loading.

Aynjell · Post by **Aynjell** » Mon Oct 31, 2005 5:03 am

Dunno man.

ThElInK · Post by **ThElInK** » Mon Oct 31, 2005 6:39 am

Aynjell wrote:Dunno man.

Well, guess I'm back on Windows to play games then. Pity.

ThElInK · Post by **ThElInK** » Mon Oct 31, 2005 7:43 pm

I tested doom3 again today, and this time when the game froze the monitor turned itself off. What does that mean?

Bolli · Post by **Bolli** » Tue Nov 01, 2005 7:01 am

Aynjell: With your EEEEEEEEEEEEEEEEEEEEE sound and an included hangup i think you can lay back some money for a new cpu .....

i had this a month ago changed powersupply made memtest used new memory and then made a test-change of the cpu and it worked gently ..
with my nvidia inmy notebook i have sometimes problems at beginn but that doesn't matter really.

so far go get some money

hogosha · Post by **hogosha** » Tue Nov 01, 2005 10:51 pm

try running an app in a window mode, and then check dmesg occasionally. i had noticed on my nvidia card some games were throwing some commands that it simply didn't like. Also try downgrading the drivers.

Birtz · Post by **Birtz** » Tue Nov 01, 2005 11:16 pm

Have you tried new 7676-r1 nvidia-kernel

Code: Select all

# echo "media-video/nvidia-kernel ~amd64" >>/etc/portage/package.keywords
# emerge nvidia-kernel
reboot system

This is ChangeLog:

Code: Select all

*nvidia-kernel-1.0.7676-r1 (28 Oct 2005)

  28 Oct 2005; Jeremy Huddleston <eradicator@gentoo.org>
  +files/1.0.7676/NVIDIA_kernel-1.0-7676-1321905.diff,
  +files/1.0.7676/NVIDIA_kernel-1.0-7676-1359015.diff,
  +files/1.0.7676/NVIDIA_kernel-1.0-7676-1361053.diff,
  +files/1.0.7676/NVIDIA_kernel-1.0-7676-1386866.diff,
  +files/1.0.7676/NVIDIA_kernel-1.0-7676-1392404.diff,
  +nvidia-kernel-1.0.7676-r1.ebuild:
  Patches for some amd64 issues from upstream.

It did the trick for me (Xid problems with some games).

ThElInK · Post by **ThElInK** » Thu Nov 03, 2005 9:22 pm

I'm already using the 7676-r1 nvidia-kernel.

I tried running doom III in window mode, and it still crashes. However, this time I was SSHed into the machine. As soon as Doom III crashed, I started getting the following in /var/log/messages:

Nov 3 22:04:47 silver sshd(pam_unix)[13397]: session opened for user root by root(uid=0)
Nov 3 22:05:06 silver NVRM: Xid: 25, L0 -> L0
Nov 3 22:05:06 silver NVRM: Xid: 6, PE0000 1f08 00000000 00000000 00f1efeb 00000000
Nov 3 22:05:09 silver NVRM: Xid: 8, Channel 00000020
Nov 3 22:05:17 silver NVRM: Xid: 8, Channel 00000020
Nov 3 22:05:25 silver NVRM: Xid: 8, Channel 00000020
Nov 3 22:05:33 silver NVRM: Xid: 8, Channel 00000020
Nov 3 22:05:41 silver NVRM: Xid: 8, Channel 00000020
Nov 3 22:05:49 silver NVRM: Xid: 8, Channel 00000020
etc...

dmesg shows:

NVRM: Xid: 8, Channel 00000020
NVRM: Xid: 8, Channel 00000020
NVRM: Xid: 8, Channel 00000020
NVRM: Xid: 8, Channel 00000020
NVRM: Xid: 8, Channel 00000020
NVRM: Xid: 8, Channel 00000020
NVRM: Xid: 8, Channel 00000020

When I kill doom 3 via ssh, the process seems to die, but I still see doom III on screen and everything looks frozen - but it isn't - SSH works, albeit very slow (you actually have to wait for a character to appear on screen after you type it).

I can't downgrade to the previous drivers, because they won't compile, giving an error I forgot.

Grooby · Post by **Grooby** » Sat Nov 19, 2005 6:03 am

I had something like this but it is regarding to lockup of my X process. It was working fine til I had a power outtage at my house. I then turn the PC off over ngiht and it's now fixed again. (another thing to check is the temperature of the GPU)

LubosD · Post by **LubosD** » Sat Nov 19, 2005 2:21 pm

NVRM: Xid is a typical nVidia problem...hard to solve

But you can try leaked 8168 nVidia driver... warez is forbidden here so I (probably) can't write here where to get it, can I?

Aynjell · Post by **Aynjell** » Sun Nov 20, 2005 2:47 am

Bolli wrote:Aynjell: With your EEEEEEEEEEEEEEEEEEEEE sound and an included hangup i think you can lay back some money for a new cpu .....

i had this a month ago changed powersupply made memtest used new memory and then made a test-change of the cpu and it worked gently ..
with my nvidia inmy notebook i have sometimes problems at beginn but that doesn't matter really.

so far go get some money

Said noise is emitted from teh on unit speaker built into the video card. It raises hell when you have said card in an AGP slot on a running machine if it's 4 pin molex isn't filled with lots of juicy 12v goodness.

My CPU is working wonderfully right now.

ThElInK · Post by **ThElInK** » Mon Nov 21, 2005 3:51 pm

LubosD wrote:NVRM: Xid is a typical nVidia problem...hard to solve

But you can try leaked 8168 nVidia driver... warez is forbidden here so I (probably) can't write here where to get it, can I?

I would rather stick with the driver in portage... Besides, the nvidia bug report is blaming the kernel, saying that kernel 2.6.14 supposedly solves the problem.

BTW, I installed SUSE 10.0 on another hard disk in this PC, and used the latest nvidia drivers available via yast, in combination with the 10.0 default kernel (2.6.13-5). The same exact problems I have on Gentoo are present in SUSE.