Berkeley AI Research · 2026-07-29模型评分46
从 CUDA 到 MLX：K-Search 如何把数十年内核经验带到 Apple Silicon
From CUDA to MLX: How K-Search Brings Decades of Kernel Expertise to Apple Silicon
导读
伯克利 Sky Lab 团队为 K-Search 增加 MLX 后端，并开发结构化 CUDA-to-MLX 翻译层，让这一进化式内核搜索框架把已有 CUDA 内核当作知识库，自动改写为 Apple Silicon 上的高质量内核。
分类置信度
高
- 官方或准官方来源，且无冲突类别。
- 正文、可信日期、主分类和证据摘录完整。
来源级别：官方一手
事件类型：研究发布
证据摘录
- “We built on K-Search , an evolutionary kernel search framework introduced by Cao et al. at Berkeley Sky Lab that uses AI to optimize GPU kernels, and extended it with a backend for MLX — Apple’s machine-learning framework for its own Apple Silicon chips.”
- “The MLX backend is built on top of the open-source K-Search repo, so the results here can be reproduced directly.”
- “We do not yet know how far this generalizes, but the result is encouraging.”
来源：Berkeley AI Research · bair.berkeley.edu