Let's say that I create my geometry, and I want to make a part of it moving. Hence, i have to create the mesh taking into account the whole path of the object. Par exampe my object is 5cm in X-dir, it moves for 10cm in X-dir, so the mesh dimension in X-dir should be 15cm. So, in the first time step, the 10 out of 15cm will be totally empty.

My question is this: do the empty cells of the mesh affect the calculation time? Do I have to create "blocking boxes" so as to cover all the empty cells each time step?

In this case you will need a big mesh. If you use "blocking boxes" then you will have to make them moving so that part of interest can move into new location. For a moving object all the cells necessary are so all the cells occupied by moving object are allocated to the memory. So using the "blocking boxes" will not help much in this case. You can use a stationary blocking box in the part of the domain where the moving object or any fluid won't arrive.

Now about the "Empty cells": all the cells not occupied by a solid object (open cells) get allocated to the memory. But if a open cell is empty then the solver doesnot spend time performing calculations in this cell. So empty cells will affect the calculation

time as they occupy memory but the solver doesnot get completely bogged down as it doesn't have to spend time in calculation in

these cells. In general it is recommended to block as many cells as possible.

looks like the last line got truncated in my earlier response. Here is the correct line: So empty cells will affect the calculation time as they occupy memory but the solver doesnot get completely bogged down as it doesn't have to spend time performing calculations in these cells. In general it is recommended to block as many cells as possible.