This project focuses on the simulation and analysis of synchronization impairments in distributed massive MIMO systems, with a particular emphasis on user-centric architectures and power control strategies. In realistic deployments, differences in local oscillators and hardware among distributed Access Points (APs) introduce synchronization errors such as carrier frequency offset (CFO), sampling frequency offset (SFO), and timing misalignment. We aim to model these impairments and develop AI-assisted synchronization schemes that can dynamically mitigate their impact. GPU acceleration is critical for this work, as the simulations involve computationally intensive signal processing (e.g., FFTs, matrix operations) and training of neural networks for adaptive synchronization and power allocation. By leveraging Alvis' modern NVIDIA GPUs, we will evaluate performance under realistic channel models and topologies, contributing insights toward the design of scalable and robust synchronization mechanisms for future 6G systems.