Our current approach for constructing interpolation of a numerical field works only in serial cases, i.e., the B field is stored in one piece in memory. How can we extend it to support interpolating of a large field that does not fit in the memory of one core/node?
Our current approach for constructing interpolation of a numerical field works only in serial cases, i.e., the B field is stored in one piece in memory. How can we extend it to support interpolating of a large field that does not fit in the memory of one core/node?