{
  "id": 312442,
  "title": "From CUDA to MLX: How K-Search Brings Decades of Kernel Expertise to Apple Silicon",
  "url": "https://urgent.news/2026/07/29/from-cuda-to-mlx-how-k-search-brings-decades-of-kernel-expertise-to",
  "topic": "tech",
  "section": "Tech",
  "published": "2026-07-29T09:00:00.000Z",
  "source": {
    "name": "Berkeley AI Research",
    "slug": "berkeley-ai-research",
    "url": "http://bair.berkeley.edu/blog/2026/07/29/cuda-to-mlx-k-search/"
  },
  "original_language": "en",
  "account": "From CUDA to MLX: How K-Search Brings Decades of Kernel Expertise to Apple Silicon\n\nThe CUDA ecosystem has amassed decades of kernel optimization knowledge, but transferring that expertise to new hardware, such as Apple Silicon, is challenging. K-Search, an evolutionary kernel optimization framework developed by Shiyi Cao at UC Berkeley Sky Lab, aims to bridge this gap. By combining AI-driven optimization with Apple's MLX framework, K-Search can adapt existing CUDA kernels for high-performance execution on Apple Silicon chips.",
  "summary": "Figure 1: CUDA-to-MLX optimization translation map. CUDA optimization knowledge can be translated into architecture-native MLX strategies rather than copied instruction-for-instruction. We face a new epoch in computing. Hardware is changing rapidly — not just faster GPUs, but a growing range of chips from different vendors, each with its own architecture and often tailored to specific AI…",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}