KernelFoundry: 带硬件意识的进化GPU内核优化
KernelFoundry: Hardware-aware evolutionary GPU kernel optimization
AI总结：
KernelFoundry通过MAP-Elites搜索、元提示进化和模板参数优化，高效探索GPU内核设计空间，实现SYCL内核在KernelBench上的2.3倍加速。
摘要：
优化GPU内核对大语言模型（LLMs）而言比标准代码生成更具挑战性，因其需理解硬件架构、并行优化策略和性能分析结果。现有LLM方法多依赖简单提示和反馈循环，仅间接通过性能分析反馈纳入硬件意识。我们提出KernelFoundry，一种进化框架，通过三个机制高效探索GPU内核设计空间：（1）MAP-Elites质量-多样性搜索，结合内核特定行为维度，维持在多样化优化策略上的探索；（2）元提示进化，协同进化提示与内核以揭示任务特定优化策略；（3）基于模板的参数优化，调整内核以适应输入和硬件。我们在KernelBench、robust-kbench和定制任务上评估该框架，生成SYCL内核作为跨平台GPU编程模型，并生成CUDA内核与先前工作对比。我们的方法在KernelBench上实现SYCL内核的平均2.3倍加速。此外，KernelFoundry作为分布式框架，支持远程访问多种硬件，实现快速基准测试，具备灵活的用户输入层，支持广泛的实际应用场景中的内核生成。
英文摘要：
GPU kernel optimization challenges LLMs beyond standard coding tasks, as it requires an understanding of hardware architecture, parallel computing optimization strategies, and profiling outputs. However, most existing approaches leveraging LLMs for kernel generation apply standard prompting and feedback loops, considering hardware only through profiling feedback. We introduce KernelFoundry, an evolutionary framework that efficiently explores the space of GPU kernels through (1) MAP-Elites quality diversity search with kernel-specific behavioral dimensions to sustain exploration; (2) meta-prompt evolution that co-evolves prompts with kernels to uncover task-specific optimization strategies, and (3) a template-based parameter optimization approach to tune kernels to inputs and hardware. We evaluate this framework on Kernel-Bench, robust-kbench and custom tasks, generating SYCL kernels as a cross-platform GPU programming paradigm, and CUDA kernels for comparison to prior work. Our approach consistently outperforms the baseline methods and achieves an average speedup of 2.3 on KernelBench for SYCL. Moreover, KernelFoundry is implemented as a distributed framework with remote access to diverse hardware, allowing quick benchmarking and featuring a flexible user input layer to support kernel generation for a wide range of real use cases beyond benchmarking.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
