Start by using HeatEqn 0 as template.
Then inital problem values are in the 3D paper.
replace the evolve section of the code with call to bds.
assume periodic boundary conditions
s should need 3-4 ghost cells.
Will need to add multifab to hold all the u,v,w MAC velocities.
For the fortran allocate command, will need to pass in a tmp multifab.
(Can't be allocated on the fly)
A big part of porting the code will be to figure out which kind of box is needed for each set of nested for loops. For example, -1 to +1 is cell-centered with 1 ghost cell -1 to +2 is nodal with 1 ghost cell
s_old will likely need 4 ghost cells