Based on a discussion I followed in JWF facebook group.
Basically, render speed depends on amount of calculations the software needs to do. To show how it works, I will perform two experiments. Each experiment consist of a serie of examples which produce visually same output, but are built in different ways.
In this experiment, we will see what happens when we:
mix up many transforms (variations)
add many linked transforms
add many pre_transforms or post_transforms
Params: Apo - Chaotica
example 1: default parameters with linked waves2
DE speed: 58.3
example 2: waves2 replaced with 10 waves2 transforms with same setup (I set the amount to 0.1 so they add up to 1 and produce same effect)
DE speed: 21.1
example 3: 10 linked linears added to the waves2 transform
DE speed: 14.4
example 4: 10 linear pre_affines are added to the waves2 transform
DE speed: 40.9