High computational efficiency of the MPI-ESM code is needed to enable throughput rates of the order of several simulated centuries per day which are essential for the success of the PalMod project. Therefore, we will work on existing bottlenecks in each component of MPI-ESM and develop new or adjusted algorithms in order to improve the overall performance. Besides suitable numerical schemes, also new hardware architectures shall be investigated whether the potential for a performance gain is given.