[emmakateh]

I am working on a UDF code for magnetic field and in order to test the code quicky without waiting for initialization in Fluent, I have a version without the face and cell loops to run in VSC.

They are exactly the same but the code going into fluent returns zeros. I have reduced it to very simple equations to ensure that the face and cell loops are not the problem.

I don't think it is the code itself because the VSC version works perfectly. I think its a memory thing but I cant figure out how to change the memory within fluent.

I do have parts of my code that use factorial and pochhammer equations so the numbers get very large, but it works when going through VSC.

If anyone has any advice, please let me know!

[AlexanderZ]

did you hook functions before running calculations?
you should show code to get answer regarding your code

[emmakateh]

yes I make sure to hook the functions before initializing. The code will run when I simplify it down to just a coordinate shift so I know that it CAN run, ex:

thread_loop_c(t,domain) /* loops over all cell threads in domain */
{
begin_c_loop_all(c,t) /* loop over cells in a thread to get information stored in cells */
{
C_CENTROID(pos,c,t);

x = pos[0] + r + R; /* shift coordinates to match equation coordinates */
y = pos[1];
z1 = pos[2] - l; /* shift coordinates to match equation coordinates */

C_UDSI(c,t,0) = x;

}
end_c_loop_all(c, t) /* END: Looping over all cells */
}

but if it is more complex it doesnt run properly. I just tried increasing the solver processes used from 1 to 2 and it was able to run and give non zero results. unfortunately the results are 100% error. The results from Fluent are twice as large as the results coming from my code in VSC.

I am running the code using 3 solver processes to test