NVIDIA Green Contexts：控制 GPU 工作负载资源共享
Control How Your GPU Shares Work with Green Contexts
NVIDIA 介绍 Green Contexts，用于控制同一进程内多个独立 GPU 组件之间的资源共享，例如延迟敏感算子与吞吐型后台 kernel、数据预处理与模型推理等场景。单个 GPU 上多阶段工作流并存时，资源分配一直难以精细控制。
GPU applications increasingly consist of multiple independent components running at the same time within a single process: a latency-sensitive operator...
GPU applications increasingly consist of multiple independent components running at the same time within a single process: a latency-sensitive operator alongside a throughput-oriented background kernel; a data preprocessing stage alongside model inference; or multiple stages of a processing workflow sharing a single GPU. Controlling how GPU resources are shared between them remains difficult.
来源：NVIDIA Developer Blog · developer.nvidia.com