This is a newbie question I know, but anyway: I have been working on my new game project, and I was wondering something. When I had my surfaces in system memory, and then blitted from them to the backbuffer in video memory, it slowed the framerate a little (about 33 fps). I can understand this. When I put my surfaces into video memory, I got around 75-80 fps. This is fine. But, when I put the primary surface and backbuffer, along with all other surfaces into system memory, I got 250 fps. So, if I can get a framerate this fast by putting the primary surface and backbuffer into system memory, then what is the purpose of even using video memory? The only advantage I can see is that you don''t need to use up any RAM on surfaces. But, other than that, what are the real benefits of using video memory rather than system? I figure, if most people are putting their primary surface into video memory, then that must have some advantage over putting it in system memory (with all surfaces also in system memory). But I can''t find any advantages. So, what I''m asking basically is, is it alright to store the primary surface, backbuffer, and offscreen surfaces in system memory? Or should I always put the primary surface and backbuffer in video memory?

I''m not entirely sure of what you did, but you really shouldn''t be able to put you primary surface in system memory. The primary surface represents the memory actually read to send to the monitor.

The back-buffer in system memory on the other hand, could make such a difference, it really depends on the speed of your video cards blitter, the speed the CPU can write to the video memory, and the amout of overdraw the back-buffer receives.