To enable deep learning models to perform Compute Unified Device Architecture (CUDA) computations on a Deep Learning Virtual Machine (DLVM), identifying the correct hardware configuration is crucial. What component is essential for configuring the DLVM to support CUDA processing effectively?