Nvidia GLX causes hard crash

[impatient]

Heres my setup.

GF 6600 GT
AMD 2500+
NForce2
1 Gb ram

I'm attemping to install Nvidia drivers 1.0.6629, however when I run glxgears , after about 5 seconds the box crashes and requiring that I power cycle. (It does not respond to pings or attempted ssh logins)

What I've tried:

• opengl-update nvidia -> Crashes after about 5 seconds
  
• opengl-update xorg-x11 -> Crashes after about 5 seconds
  
• opengl-update xfree -> Works fine, but it's horrribly slow
  

As implied by above I have tried both the latest Xfree and Xorg xservers, versions 4.3.0-r8 and 6.8.0-r3 respectively.

The nv-control extension seems to load correctly as well as well as the nvidia glx extension.


The Xfree logs do not appear to show anything useful. I can attach them if necessary.

I've recompiled my kernel a couple times as well... I did a make old config for the new kernel(2.6.9) which created a couple of problems such as disabling devfs and changing my processor type.(How did that happen)

The nvidia drivers worked fine with the old video card, but those were a couple versions old.

My x config files have

[d_m]

Probably a lot of junk in there, but you should post more of xorg.conf. In particular, the video card section where the "nvidia" driver is being loaded.

I've never used the opengl-update command before--in my experience just loading glx and using the "nvidia" driver, then restarting xdm are all I do. Maybe someone more experienced can help.

What happens if you start X (/etc/init.d/xdm start) with "Driver 'nvidia'" in the video card section? X dies after 5 seconds??

[EDIT: also, what do var/log/Xorg.0 etc. say?? /var/log/messages?? ]

[impatient]

I can play quake3 at lower resolutions ok, but glxgears still crashes after 5 seconds. The second I go beyond 1280x1024, everything breaks. I'm afraid that when I really tax the box that everythings going to continuosly lock up.

Here is my config file. The log can be found at http://derivate.us/xfreelog.txt

And here is the config file

[mike4148]

In my experience, system lockups involving the nvidia drivers usually have something to do with either the version of the drivers or AGP. Try another version of the drivers, if you can; otherwise, mess around with your AGP setup. To start with, read the readme at /usr/share/doc/nvidia-glx-*/README.gz. Try building agpgart support for your chipset into your kernel (or load it as a module) and use NvAGP="2" in the "Device" section for your card in your xorg.conf. If that doesn't work, try NvAGP="0" to disable AGP. If everything magically works (but is probably very slow), see if your BIOS has any settings you can play with. Beyond that, seek ye the many other threads about solving AGP issues with the nvidia driver on the forum system.

[impatient]

I tried everything but the generic kernel module. (Which seems pointless if it crashes sans agp.)

The NvAGP 2 option seemed to work and I was able to play Q3 for about 10 minutes maxed out... of course, then it crashed.

The NVAGP 0 option looked like it was going to work ok(there were some artifacts/misdrawn items), but that crashed more quickly than NvAGP 2.

The older drivers won't work.. all I get is a black screen.

My main concern right now is that I may have bad hardware. I guess I can install windows and try it out.

I thought of changing the os-registry.c file, but since it crashed with no agp this probably wouldn't solve the problem.

(I also looked at a lot of the other posts out there... to no avail)

Thanks for the suggestions. Any more possibly?