Beamforming using Discrete Fourier Transform (DFT) based codebook is widely studied for millimeter wave (mmWave) communication systems. This design requires only phase-control for each antenna element and is therefore motivated by the possibility of lower complexity hardware. While this DFT-based design works well for single-user transmission, significant inter-beam interference is generated in multi-user multiple-input multiple-output (MU-MIMO) transmission. Recently, beamforming based on amplitude tapering has been demonstrated even for mmWave systems. The non-uniform amplitude in this design allows the possibility of significantly reducing inter-beam interference at the cost of slightly reducing the main lobe gain. Since the amplitude tapering is fixed and designed offline, the additional implementation complexity is not very high. In this paper, we show that Dolph-Chebyshev and Taylor codebook designs can provide significant improvement in performance over DFT-based codebooks in full-dimension MU-MIMO settings. For the MU-MIMO, we also propose a per-user power allocation algorithm for maximizing the sum rate under total power and rate constraints. The results show that the proposed algorithm gives high sum rates compared to equal power allocation among users. The simulations are carried out under the 3GPP full-dimension MIMO channel model. © 2020 IEEE.