Loop over cells per processor (parallel)
 

Hi I got a problem and I have read a lot of posts on a different forums but still could not find a solution. I am doing something wrong. Could some one help me out.

Basically when I run my program on single processor it works fine but when on parallel it does not.
I am using the loop

forAll(mesh.C(),celli){

x = mesh.C()[celli].component(0);
y = mesh.C()[celli].component(1);
z = mesh.C()[celli].component(2);
arr = getnewvelocity(rho, U[celli].x(), U[celli].y(), U[celli].z(), p[celli]);

// .......

}

I found a post on this forum suggesting to use "reduce" command but it means that I send all data from all proc to one. I dont want this. I wan to go through all cells per processor.

Could some one give me an advise or example.
Thank you

Well you can. But you must realize that in parallel this only goes through all the cells that are on this processors. But on all processors. Which makes it alright

Whether you want to do a reduce or not is irrelevant: it depends on what is happening in the location where you put the 7 points. May be that you don't need a reduce. May be that you need it after the loop.
That is hard without knowing what the 7 points do