OpenGL performance hints
The reason this is faster is the display list can optimize this type of
primitive into a single, very efficient structure. The small cost of adding
extra data is offset by this optimization.
Performance is increased by maximizing the number of vertices per
Begin/End pair. If your vertex data in memory is organized in a linear,
rather than a random manner, performance is enhanced by taking
advantage of vertex pre-fetch. It is most efficient to use 32-bit float data,
which avoids the need to convert data.
texture downloading performance
This section includes some helpful hints for improving the performance
of your program when downloading textures.
If you are downloading MIP maps, always begin with the base level
(level 0) first.
If it is possible, you should use texture objects to store and bind
If you are doing dynamic downloading of texture maps, you will get
better performance by replacing the current texture with a texture of
the same width, height, border size, and format. This should be done
instead of deleting the old texture and creating a new one.
To increase the performance of selection (glRenderMode GL_SELECTION)
it is recommended that the following capabilities be disabled before
entering the selection mode.