paper-with-me

홈 › Papers

Optimizing CUDA like a Human: Micro-Profiling Tools as Expert Surrogates for LLM-Based GPU Kernel Optimization

2026-06-24 · Jiading Gai, Shuai Zhang, Kaj Bostrom, Jin Huang, Vihang Patil, Haoyang Fang, Bernie Wang, Huzefa Rangwala, George Karypis arxiv

We present KernelPro, a closed-loop multi-agent system that automatically generates, profiles, and iteratively optimizes GPU kernel code by integrating large language model (LLM) code generation with hardware profiler feedback and pluggable bottleneck detection tools. KernelPro introduces four contributions: (1) a semantic feedback operator that encodes expert heuristics as pluggable micro-profiling tools, transforming raw hardware metrics into actionable natural language guidance; (2) a two-stage tool invocation architecture where roofline-based bottleneck classification filters which specialized analysis tools execute, combining kernel-level (ncu), instruction-level (SASS), and system-level (nsys) profiling; (3) a domain-adapted MCTS with progressive widening, asymmetric branching, log-reward calibration, dead-end pruning, and search memory for cross-iteration learning; and (4) direct CuTe source-level code generation via autonomous code search over the CUTLASS/CuTe codebase. On KernelBench, KernelPro achieves geometric mean speedups of 2.42x/4.69x/5.30x on Levels 1/2/3, establishing state-of-the-art performance across all difficulty levels. On VeOmni's expert-optimized MoE training kernels, KernelPro achieves 1.23x over hand-tuned Triton by generating a from-scratch raw-CUDA+CuTe Hopper WGMMA kernel. Ablation studies demonstrate that each design component independently and significantly improves optimization quality: micro-profiling tools (p < 0.0001 vs raw metrics), MCTS search (26% higher geometric mean vs greedy, p = 0.004), and proactive tool orchestration (23% improvement, p = 0.035). Finally, KernelPro is the first CUDA kernel coding agent to optimize energy efficiency beyond the speed-only focus of prior systems, demonstrating an 11.6% measured energy reduction at matched speed.

📄 PDF Abstract BibTeX arXiv:2606.26453

Code (0)

등록된 구현이 없습니다.

Tasks

Code GenerationCode Search

Similar Papers 제목 키워드 기반

Making LLMs Optimize Multi-Scenario CUDA Kernels Like Experts

2026-03-07 · Yuxuan Han, Meng-Hao Guo, Zhengning Liu, Wenguang Chen 외 arxiv

Optimizing GPU kernels manually is a challenging and time-consuming task. With the rapid development of LLMs, automated GPU kernel optimization is gradually becoming a tangible reality. However, current LLM-driven automa…

CUDA Agent: Large-Scale Agentic RL for High-Performance CUDA Kernel Generation

2026-02-27 · Weinan Dai, Hanlin Wu, Qiying Yu, Huan-ang Gao 외 arxiv

GPU kernel optimization is fundamental to modern deep learning but remains a highly specialized task requiring deep hardware expertise. Despite strong performance in general programming, large language models (LLMs) rema…

Reinforcement LearningCode Generation

MAP: Microblogging Assisted Profiling of TV Shows

2015-02-11 · Lin Xiahong, Wang Zhi, Sun Lifeng

Online microblogging services that have been increasingly used by people to share and exchange information, have emerged as a promising way to profiling multimedia contents, in a sense to provide users a socialized abstr…

Retrieval

AscendOptimizer: Episodic Agent for Ascend NPU Operator Optimization

2026-03-24 · Jiehao Wu, Zixiao Huang, Wenhao Li, Chuyun Shen 외 arxiv

Optimizing AscendC (Ascend C) operators for Ascend NPUs is difficult for two reasons. First, unlike CUDA, the ecosystem offers few public kernels to learn from. Second, performance depends on a coupled two-part implement…

Optimizing Block-Sparse Matrix Multiplications on CUDA with TVM

2020-07-26 · Zijing Gu

We implemented and optimized matrix multiplications between dense and block-sparse matrices on CUDA. We leveraged TVM, a deep learning compiler, to explore the schedule space of the operation and generate efficient CUDA …