hi, I got this funny problem maybe someone might know something about it, my LFS system was recently built and everything is working fine, but when im trying to use SDL 1.2.8 the rendering of video data is very slow, as in slower than software mode (e.g. 1 FPS or less) are there any options during SDL configuration I should have used but was supposed to ? the BLFS book version 6.0 is not yet complete but it does not seem to have any special options, so I think it should have worked, the strange thing is that when I use the GLX rendering mode it is also just as slow, if the SDL has anything to do with GLX or X then my video drivers are ok and working but not in the LFS system, would anyone know what is happening ?

but 2D acceleration is working fine, and all software rendering as well

as far as I know openGL is fully supported by the nvidia drivers (they wouldn't be of any use otherwise)

ooopsssss............sorry I think this information should help, haven't noticed that and I think it has a lot to do with the problem :

I *think* the "forceSW" is the problem, but I haven't done anything to trigger it therefore I don't know how to turn it off, would you know how to fix it ?

my guess is that the "forceSW" means "force software" as for how and why it was turned on I have no idea :/

the funny thing about it is that it worked before, and now it doesn't, so I started thinking what would cause it to simply stop working
after finding out about the above I also looked into 2 other things :

I had "options nvidia NVreg_ReqAGPFW=1" in the modprobe.conf file
which I recently disabled, but it did not help.

I have also tried other version of the driver, but none of them seem to work with both fluxbox and 2.6.10
so this is the only driver, but it should work so now I know it is not a driver problem for sure, the rest of the problem is this :

in addition to the above information, I have also installed an another network card recently into the system
apperantly one of them is sharing an interrupt with the video card, it is true that before all the interrupts had their own
channel so I think this could be it, my BIOS supports assigning IRQ to pci slots individully but my kernel is configured
to use direct pci access, should this really be the problem I believe it would fix it, if not I hope to find a good
pci configuration by moving the cards around (not that there is much to move as you may have noticed), but in the former case
I will need to recompile the kernel with BIOS PCI access I believe, assuming it would work, what is the difference between direct
pci access and bios ? and why would there be an interrupt problem if I have APIC ? APIC does allow sharing IRQ safely does it not ?

oh and while we're on this subject, what do all those additional digits in the 16 and higher interrupt range ? and the 4 3-letter names at the bottom what are those ?

and last but not least..........I don't have an SMP system, I know that io-apic allows assigning an interrupt per cpu some sorts of irq-routing, but does it have any meaning for single-processor system ? would a non-io apic and an io-apic make any difference for a single-cpu system ?

does anyone know why this is happening ?