苹果 Minions：闲置 Mac 组队成虚拟 GPU 云
苹果研究团队发表论文，介绍内部系统 Minions：它把网络中闲置的 Apple Silicon Mac 聚合成一个动态调度的“虚拟 GPU 云”，用于执行机器学习推理任务。系统由 Minion Server 拆分任务、PostgreSQL 保存任务元数据，各台 Mac 上的 Minion Executor 通过 REST 接口主动拉取子任务并在本地 GPU 运行、从 Amazon S3 读写数据，Minion Watchdog 依据每五分钟一次的心跳把休眠、离线或关机设备的未完成任务重新排队。论文称目前只处理推理，任务类型包括 Hugging Face Transformer 模型的向量嵌入和通过各机预装 Ollama 进行的文本生成；苹果自述部分原本需在集中式 GPU 队列等待逾 10 小时的推理任务，在 Minions 下数分钟内即可开始执行，但该数据来自苹果自身测试，且跨团队数据保密仍是顾虑。
苹果在论文中表示，在集中式 GPU 资源受限的环境里，闲置的 macOS 设备可以被转化为实用的分布式 GPU 推理平台。
research area Speech and Natural Language Processing, research area Tools, Platforms, Frameworks
content type paperpublished October 2026
Minions: A Dynamic and Distributed Virtual GPU Cloud Built on Idle Apple Silicon Devices
AuthorsJianhua Li, Yun Xing, Girisha De Silva, Ryan Flint, Kai Yuan, Sean Suchter
The increasing demand for GPU-accelerated machine learning inference has exposed limitations in centralized GPU infrastructures, including high cost, limited elasticity, and long wait times under resource constraints. At the same time, modern Apple Silicon–based macOS devices are equipped with powerful GPUs that remain largely idle during normal usage. This paper presents Minions, a dynamic and distributed virtual GPU cloud that aggregates idle Apple Silicon GPUs across macOS devices to execute large-scale machine learning inference workloads. Minions enables teams to leverage existing hardware without requiring additional GPU procurement. The system provides flexible GPU inference task execution, fault tolerance, and efficient resource utilization. Our experience shows that idle macOS devices can be transformed into a practical distributed GPU inference platform in environments where centralized GPU resources are constrained.
#Apple#Minions#Apple Silicon#Mac#Ollama#Hugging Face#PostgreSQL#Amazon S3