Kernel Foundry: A Diagnosis-driven Evolutionary Kernel Optimizer with Multi-Experts
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2026
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866911729190961152 |
|---|---|
| author | Huang, Zixuan Chen, Da Huang, Kecheng Yin, Lihao Li, Xing Zhen, Huiling Yuan, Mingxuan Shao, Zili |
| author_facet | Huang, Zixuan Chen, Da Huang, Kecheng Yin, Lihao Li, Xing Zhen, Huiling Yuan, Mingxuan Shao, Zili |
| contents | Generating high-performance GPU kernels remains challenging due to the need for both correctness and hardware-aware optimization. While large language models (LLMs) show promise in code generation, they often fail to produce kernels that are both correct and efficient.
We propose Kernel Foundry, a diagnosis-driven evolutionary framework for automatic GPU kernel optimization. Our method combines expert-guided, retrieval-augmented initialization with a multi-island evolutionary search, where candidate kernels are iteratively refined using structured diagnostic feedback. A centralized experience library accumulates reusable optimization knowledge to guide subsequent evolution, while explicit mechanisms prevent cheating behaviors that bypass kernel-level computation.
Experiments on KernelBench show that our method consistently improves both correctness and performance over strong baselines, achieving up to 100% correctness on Level~2. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_30359 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Kernel Foundry: A Diagnosis-driven Evolutionary Kernel Optimizer with Multi-Experts Huang, Zixuan Chen, Da Huang, Kecheng Yin, Lihao Li, Xing Zhen, Huiling Yuan, Mingxuan Shao, Zili Neural and Evolutionary Computing Distributed, Parallel, and Cluster Computing Machine Learning Performance Software Engineering Systems and Control Generating high-performance GPU kernels remains challenging due to the need for both correctness and hardware-aware optimization. While large language models (LLMs) show promise in code generation, they often fail to produce kernels that are both correct and efficient. We propose Kernel Foundry, a diagnosis-driven evolutionary framework for automatic GPU kernel optimization. Our method combines expert-guided, retrieval-augmented initialization with a multi-island evolutionary search, where candidate kernels are iteratively refined using structured diagnostic feedback. A centralized experience library accumulates reusable optimization knowledge to guide subsequent evolution, while explicit mechanisms prevent cheating behaviors that bypass kernel-level computation. Experiments on KernelBench show that our method consistently improves both correctness and performance over strong baselines, achieving up to 100% correctness on Level~2. |
| title | Kernel Foundry: A Diagnosis-driven Evolutionary Kernel Optimizer with Multi-Experts |
| topic | Neural and Evolutionary Computing Distributed, Parallel, and Cluster Computing Machine Learning Performance Software Engineering Systems and Control |
| url | https://arxiv.org/abs/2605.30359 |