Based on OpenGL programming guide to cache data, for example, OpenGL can discard cached data directly, improve efficiency, I'm not very understand, this is discarded release this section of memory, or other operations, abandon away part of the data, change the cache size?
Such as I have a BO as VAO vertex data source, which contains 4 * 4 points, so I apply colours to a drawing is completed, the next frame need only the first four points in the cache, then discard redundant vertex data, then again the next frame and I want to 4 * 4 point data, at this moment is directly using the original BO, or use bufferdata functions to apply for memory?