Is it any difference between using the glNewList function to collect display functions and glCallList to call the list,
or just defininig simple function that consists of the same callings and calling the function instead of calling the glCallList?

I’m not sure exactly what you’re asking, but if you’re asking “Is there a performance difference between using display lists and not using display lists”, then the answer is usually “Yes!”.

You see the largest performance improvements when you have many vertices or gl functions inside the display list, otherwise the function call of the list can be as/more expensive than the few gl functions. For example, putting the vertices for a cube into a display list doesn’t speed things up much if at all, but putting the vertices to a large model into a display list has improved the performance in my demos by 300% or more.

Hope that helps.

It is true that the use of dlist will improve largely the efficiency, but if it double also the memory? Suppose we have a very large finite element model composed by million tetrahedrons, can still use display list in this case?