Partial fix of double grid
Partially fixes the domain parallelization of the double grid feature. When running the corresponding test in parallel, the discrepancy in total energy between 1 and 3 procs has been reduced from 1.1e-1 Ha to 9e-5 Ha. The code now converge in both cases in the same number of iterations (compared to 10 and 7 before). Still this is not working properly...
Some fixes for DoubleGrid in domain parallelization.
- I have checked that my code follows the Octopus coding standards
- I have added tests for all the new features added in this request.