slow transformation in For loop
Hello gmsh developers.
I noticed some strange behavior. I want to duplicate and move an object in a for loop. The cpu time is not linear. It is quadratic as a function of the number of duplications.
I read in the docs that this is normal with the built-in geometry kernel because of the autocoherence setting. But in my case I use the opencascade kernel.
I have attached a simple example of duplication / translation of a box. The box is duplicated 100 times. On my machine it takes 1.7 sec. If I duplicate it 400 times it takes about 23 sec, which is about 13.5 times longer. It is close to a quadratic behavior.
If I replace the duplication / translation by the creation of a new box, we find a linear behavior. But my need is to duplicate and place complex objects.
Is there a bypass to find a linear behavior or is it a constraint of the geometric kernel?
Thank you very much.untitled.geo