Up to now I have accepted that the random lockups while running Vo were due to memory leaks from plug-ins or the video card overheating. Due to recent improvements in my ability to monitor my gpu and the rework of several plug-ins I realised this was not the case. To eliminate Plug-ins from the picture I unloaded them all.

Gpu temp is 54 C while Vo is running and 38 C when gpu is idle. I have a GPU cooler installed as well.

I am only using about 1 gig of memory with no signs of a memory leak with or without plug-ins

At lock up the video goes fuzzy with vertical streaks. I can move the mouse for a few seconds but it locks up soon after. All keyboard input is also locked so I cant shut my system down gracefully and must do a hard reboot.

Video card is a Nvidia Gforce 6200 with 256m of memory. Video mode is setup for 1680x1050 at 60 hz. I have VO setup for the high setting (256) with glow turned off. Frame rates in station are at 30 and 60 to 90 in space. No slowdown or artifacts are observed before Lockup.

System is a dual core Amd running Suse linux 11.1 with KDE 3 desktop and 2 gig of memory. (No swap) 1 Gig of free memory when VO is loaded and running.

Any ideas,questions or suggestions? I would like to solve this.

Seems an awful lot like a hardware failure to me. Maybe your GPU is just damaged from the prior overheating?

The first thing would be to lower the texture quality and see if that helps. Some drivers don't react well to all of the video memory being consumed.

Do you notice that it locks up if something enters your field of view or something otherwise new happens?

Uh no the card wasnt overheating... I just did not have a way to check the temp before....

And a hardware problem would show up on other applications. VO is the only one locking up. I suspect this is a video memory problem of some type but I dont know of any way to monitor the usage of it.

I havent noticed if any particular objects do it. I will try turning down the texture and get back to you. I do seem to render the bone roids with no problem....

There's also a way to turn off vertex buffers by editing the config.ini file. That may also help. It seems to be a source of misc crashes. Sometimes locking the vertex buffer returns an invalid pointer. Hmm coming to think about it, maybe there's some obscure AGP windowing bug that only manifests itself when all texture memory of the video card is exhausted or something.

possible.
I also just ran across a bug in the vsync on the card itself so i am disabling it to check its effect. Its sync to Vblank....

BTW while I am verifing the operation...

A better indicator of Video memory needed may help. Currently the only clue is when setting the standard video settings. A running estimated total for the options selected might be a better way for us to choose our options....Just a thought.

nope... still locks.

Trying a lower video setting. Changed from 1680x1050 60 hz to 1152 x 864 75 hz. The fonts look nasty so I doubt I will stay in this mode for long....

nope still crashing...

If it's an agp card, setting the agp mode to 4x and switching off fast writes in the bios setup might help.

FSAA used to also cause crashes with my old nvdia card.. but that was geforce 2 and a couple of years ago.

Try changing these values in your config.ini:

[refgl]
tc=0
doshaders=0
dovertexbuffers=0
doindexbuffers=0
do_compiled_vertex_array=0

Worst case you can try this one:
do_extensions=0

which turns off all usage of OpenGL extensions.

Well I changed back to the 1680x1050 mode and set the texture down to lowest.... no crash so far....I will try those if I need too.