Matrix for Poisson equation

[Viktor]

Hi!

I have a question regarding the way the Laplacian matrix is constructed in mantaflow for the pressure projection step. The values of the matrix as it is constructed in the "MakeLaplaceMatrix" function at the conjugategard.h file, are not scaled with (1/dx)^2 because mantaflow assumes unit grid cell size, right?

Thank you!

[thu...@gmail.com]

Hi, yes - exactly. The grid in mantaflow has a normalized cell size of dx=1, which leads to a constant Laplace stencil.