Dont know why but i cant post a reply from any computer so I am assuming that i am writing a too long message so here is the explenation of what i am working on and the next reply will be the code with comments.


I am working on stable fluids introduced by Jos Stam. Trying to implement into 3d. The representation of the velocity field is not correct in my program and I think that this part of code is not doing the job correctly. The reason I use pointer is because i have a class defined for my physics solver cube so i find it nescesary to use pointers.

link to my code cause i cant seem to post my code for some damn reason.

http://pastebin.com/vQAxpnhZ

Quote:Orignal post by tompish
The reason I use pointer is because i have a class defined for my physics solver cube so i find it nescesary to use pointers.

At the end the decision is up to you, but I don't see the reason. Somewhere the routine Semiadvect is invoked, and that invocer should be able to read the values from Cube and put them as parameters for Semiadvect. Inside Semiadvect there is no reason to change those values; you actually don't change them there, but there is no documentation about this.

Unfortunately, I don't know anything about Jos' paper. But the main questions are still there: Have you thought about the 1..N and N+2 size problems? And have you tried to drive the tri-linear interpolation isolated, to make sure it works for itself?

Do you use this paper?

Nope have not tried to isolate the interpolation, that was a good thing to do ill try that later on. And yes I have thought about the grid size.. the actual grid is 1..N while there is an extra layer of cells that account for the boundary conditions 0 and N+1.

tx for the help man.