用智能代理自动优化GPU内核，提速超6倍且省时48%。
KernelBrain: Coarse-to-Fine, Budget-Aware Search for Agentic GPU Kernel Optimization
- 分阶段搜索：先低成本筛选，再聚焦高潜力候选
- 实测最高提升6.72倍性能，优化时间减少48%
- 适合需要高效生成GPU内核的开发者与研究者
自动化GPU内核优化在实践中仍具挑战：生成变体可能违反正确性约束，运行时测量噪声大，搜索常早期停滞。我们提出一种实用的优化智能体，结合大模型引导变异、自适应资源分配、策略门控评估和剖析器驱动诊断。系统以低成本评估筛选大量候选，仅对有前景的保留者分配更高精度预算，从而优化并演化GPU内核。在关键Triton内核生成任务中，该设计显著提升内核质量和搜索效率，相较PyTorch实现0.88x-6.72x加速，相比最先进内核代理最高达1.4x加速，优化时间最多降低48%。
原文摘要 · Abstract (English)
Automating GPU kernel optimization remains difficult in practice: generated variants can violate correctness constraints, runtime measurements are noisy, and search often stalls early. We present a practical optimization agent that combines LLM-guided mutation, adaptive resource allocation, policy-gated evaluation, and profiler-informed diagnosis. The system screens many candidates with low-cost evaluation and allocates higher-fidelity budget only to promising survivors to optimize and evolve GPU kernels. On important Triton kernel generation tasks, this design improves both kernel quality and search efficiency, reaching 0.88x-6.72x speedup over PyTorch and up to 1.4x speedup over the state-of-the-art kernel agent, with up to 48% lower optimization time.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。