About Navier-Stokes Equations Solution

Hi everyone:

I’m doing the fluid project. I’m following the SPH approach and also have done a CPU version. I’m really stuck how to divide the fluid problem into pieces in order to make use of GPU parallel ability. I just consider to take each part of the equation in a kernel function to solve. Am I thinking in right way? As we know, the equation can be divide as external force, advection, pressure and viscosity. So what I’m thinking is to parallel calculate these parts.

Can anyone help me or give me a tutorial? Or if I’m doing wrong, please correct me.

“I’m doing the fluid project. I’m following the SPH approach”

i wonder how many cuda fellows are fluid in the fluid project, particularly the SPH approach
following pseudo code is of course a different story altogether

look at the common inputs/ outputs of the functions/ functionality, and the likely optimal dimensions of the functions/ functionality, in determining where best to draw functional borders perhaps
also not dependencies between functions/ functionality, to further determine if functions can be run concurrently, in addition to merely be parallelized