Except for openGL needs I use the VRAM also for saving the depth buffer (of size 512x512) and for display lists of four very heavy models.
When I load 3 models my application runs successfully. However, when I try to load 4 models I get frozen screen and stuck computer, which I need to reboot (In the frozen screen I see my application window fully transparent).
I am sure the problem is not in the models (as every one of them is loaded successfully alone).
I have 128MB of VRAM.
did my VRAM overflow?
How can I know?
or anyone of you can give me other possibilities for my computer getting stuck and my frozen screen?
The GL should not hang the computer or cease operation because of VRAM overflow.
It is not specified where, specifically, the display list is stored. AGP memory is one possible location; system memory is another. The implementation is free (and recommended) to move data around as necessary.
If you make proper API calls, and the system hangs, then there’s a bug in your GL implementation, and you should try sending the crashing program/data to the developer support people for your card to have them fix it.
You should try increasing your GART mapped memory i.e. your AGP aperture size using your PC BIOS settings (all PCs allow you to do this in the BIOS). This is the system memory that the driver can DMA from and will use this for resources like texture space and display lists when it runs out of video memory.
I had AGP aperture size of 64MB, Increasing it to 128MB or 256MB did not change the situation, I still get frozen screen and stuck computer .
I is really annoying…
Either you’re doing something wrong with ARB_buffer_region, or your implementation is buggy. You should send the driver developer a bug report, complete with a functioning example program.