Installing the GPU NVIDIA driver and CUDA Toolkit
Checking NUMA topology and affinity
Benchmarking with the CUDA demo_suite
Verifying/benchmarking the P2P capabilities
Benchmarks with Amber22 software