{"ID":2841747,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2511.13764","arxiv_id":"2511.13764","title":"Library Liberation: Competitive Performance Matmul Through Compiler-composed Nanokernels","abstract":"The rapidly evolving landscape of AI and machine learning workloads has widened the gap between high-level domain operations and efficient hardware utilization. Achieving near-peak performance still demands deep hardware expertise-experts either handcraft target-specific kernels (e.g., DeepSeek) or rely on specialized libraries (e.g., CUTLASS)-both of which add complexity and limit scalability for most ML practitioners. This paper introduces a compilation scheme that automatically generates scalable, high-performance microkernels by leveraging the MLIR dialects to bridge domain-level operations and processor capabilities. Our approach removes dependence on low-level libraries by enabling the compiler to auto-generate near-optimal code directly. At its core is a mechanism for composing nanokernels from low-level IR constructs with near-optimal register utilization, forming efficient microkernels tailored to each target. We implement this technique in an MLIR-based compiler supporting both vector and tile based CPU instructions. Experiments show that the generated nanokernels are of production-quality, and competitive with state-of-the-art microkernel libraries.","short_abstract":"The rapidly evolving landscape of AI and machine learning workloads has widened the gap between high-level domain operations and efficient hardware utilization. Achieving near-peak performance still demands deep hardware expertise-experts either handcraft target-specific kernels (e.g., DeepSeek) or rely on specialized...","url_abs":"https://arxiv.org/abs/2511.13764","url_pdf":"https://arxiv.org/pdf/2511.13764v1","authors":"[\"Arun Thangamani\",\"Md Asghar Ahmad Shahid\",\"Adam Siemieniuk\",\"Rolf Morel\",\"Renato Golin\",\"Alexander Heinecke\"]","published":"2025-11-14T14:32:28Z","proceeding":"cs.LG","tasks":"[\"cs.LG\",\"cs.PF\",\"cs.PL\",\"cs.SE\"]","methods":"[]","has_code":false}