I am trying to optimize the terrain render part of my real time game, which uses a top-down perspective.
Currently, I create all vertices in the visible area each frame and send them to the GPU (dynamic buffer).
This is costly, so I wanted to try and use a static buffer instead, where all the vertices are allocated on the GPU on startup.
However, I got an out of memory exception on my larger maps. Apparently, there were too many vertices to fit in the video memory.
Any ideas for how to move on?
I also have an idea about rendering the terrain into bitmaps and displaying them as one or more big quads.
Is this a method that is commonly used? What is the best way to go about it?