When deploying GPUs in a high-performance computing (HPC) environment, customers face substantial obstacles and inefficiencies caused by the need to port and refactor code. Their efforts are further hampered by proprietary GPU programming environments that prohibit portability between GPU vendors and often result in inconsistency between CPU and GPU implementations. The need for GPU-level memory bandwidth, at scale, and sharing code investments between CPUs and GPUs for running a majority of the workloads in a highly parallelized environment has become essential. Intel Data Center GPU Max Series is designed for breakthrough performance in data-intensive computing models used in AI and HPC.