Registration is a problem of transformation estimation between two point clouds, which has experienced a long history of development from an optimization aspect. The recent success of deep learning has vastly improved registration robustness and efficiency. This survey tries to conduct a comprehensive review and build the connection between optimization-based methods and deep learning methods, to provide further research insight. Moreover, with the recent development of 3D sensors and 3D reconstruction techniques, a new research direction also emerges to align cross-source point clouds. This survey reviews the development of cross-source point cloud registration and builds a new benchmark to evaluate the state-of-the-art registration algorithms. Besides, this survey summarizes the benchmark data sets and discusses point cloud registration applications across various domains. Finally, this survey proposes potential research directions in this rapidly growing field.