gpu-compute-management

GPU provisioning across clouds, CUDA debugging, multi-GPU coordination, spot instance strategies, and cost-aware selection.