Computationally, the calculation of computer-generated holograms is extremely expensive, and the image quality deteriorates when reconstructing three-dimensional (3D) holographic video from a point-cloud model comprising a huge number of object points. To solve these problems, we implement herein a spatiotemporal division multiplexing method on a cluster system with 13 GPUs connected by a gigabit Ethernet network. A performance evaluation indicates that the proposed method can realize a real-time holographic video of a 3D object comprising ～1,200,000 object points. These results demonstrate a clear 3D holographic video at 32.7 frames per second reconstructed from a 3D object comprising 1,064,462 object points.