NVIDIA CUDA 13.4 在 Kubernetes 中实现内存强制 GPU 共享

Memory-Enforced GPU Sharing in Kubernetes with NVIDIA CUDA 13.4

要点速览

NVIDIA CUDA 13.4 发布，在 Kubernetes 中通过内存强制机制实现 GPU 共享，允许将单个 GPU 分割为硬件隔离的实例供多个 AI 工作负载使用，提升资源利用率。

正文

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…

Source

来源: NVIDIA Technical Blog · developer.nvidia.com