An efficient continuum model for simulating polymer blends and copolymers is presented. In this model, the interactions are short-range and purely repulsive, thus allowing for excellent computational performances. The driving force for phase separation is a difference in the repulsive interaction strength between like and unlike mers. The model consists essentially of a molecular-dynamics algorithm, supplemented by an appropriate Monte Carlo exchange process. To demonstrate the effectiveness of the model we study two systems, a symmetric binary blend of polymers and a symmetric diblock copolymer system. For the binary blend, we determine the phase diagram and find, as predicted by theory, that the critical interaction parameter scales with the inverse of the chain length of the polymers. For the diblock copolymer system, we study both the one-phase region and the microphase separated lamellar region. For the latter, we show that constant-pressure algorithms are more appropriate since, contrary to recent lattice simulations, the lamellar spacing can self-adjust in such an ensemble.