Hey all, I don't know if anyone else had the question 'How does the OS share the GPU around with quartz extreme?' ... but for any who did, Andrew Welch from Ambrosia software (aka Moki) explained it nicely on a macnn forum recently.
Cheers, Tobes. quote: -------------------------------------------------- ---------------------- Originally posted by KellyHogan: Moki, with an Open GL wrapper accelerating Quartz, how much of a performance decrease can users expect in Lightwave and Maya? Logically if the graphics card is using Open GL to accelerate quite a stressful GUI that would slow down apps that needed as much Open GL performance as needed. -------------------------------------------------- ---------------------- None -- applications that are using the card in the manner that Maya and Lightwave (and games) do are given priority. The way it works can be considered analogous to the VM system that Mach uses for virtual memory. If there is room, the textures go on the video card, if not, they go in main RAM (and you get a bit of a speed hit there). The thing is, 32mb of VRAM is enough for all of the frontmost apps (such as LightWave/Maya) windows and menus to be buffered in VRAM, in addition to the texture memory those apps themselves are using up for the model textures. If you want more stuff to be buffered in your video card, you can also always toss more VRAM in it.
