[Graphics]Optimization at Deep Level
keywords: Graphics, Optimization at Deep Level
The engine draws the things you see on the screen in batches and such a batch is called a drawcall.
The time a drawcall takes consists of two parts:
- a static overhead, that is always the same no matter how many things get drawn;
- a dynamic part that depends on how many things there are to draw;
One basic thing is that each model in idTech4 gets drawn once per light that hits that model. And each of these draws of the model is split into drawcalls for each material.
So if you have a model with two different materials that is hit by 3 lights, the entire model is drawn 3 times, and the engine uses 2 * 3 = 6 drawcalls to do so.
“The three most charismatic leaders in this century inflicted more suffering on the human race than almost any trio in history: Hitler, Stalin, and Mao. What matters is not the leader's charisma. What matters is the leader's mission.” ― Peter F. Drucker, Managing the Non-Profit Organization: Principles and Practices