NVIDIA CUDA 13.4 如何在 Kubernetes 中实现内存强制的 GPU 共享
Memory-Enforced GPU Sharing in Kubernetes with NVIDIA CUDA 13.4
NVIDIA CUDA 13.4 支持在 Kubernetes 中对 GPU 进行内存强制的共享，让多个只需部分 GPU 的 AI 工作负载共用同一设备，避免整卡分配造成的容量闲置。NVIDIA GPU 此前已支持 MIG 等多种共享方式，MIG 可将 GPU 拆分为对应用透明的硬件隔离实例。
Many AI workloads need only a fraction of a GPU. Giving each one an entire device can leave capacity idle; sharing GPUs lets platform teams serve more workloads...
Many AI workloads need only a fraction of a GPU. Giving each one an entire device can leave capacity idle; sharing GPUs lets platform teams serve more workloads with the infrastructure they already have. NVIDIA GPUs already support several approaches to sharing. Multi-instance GPU (MIG) splits a GPU into hardware-isolated instances that are transparent to the application…
来源：NVIDIA Technical Blog · developer.nvidia.com