Scholarlink

Edge AI accelerators commonly use SIMT architectures, but AI kernels often suffer from high memory access overhead, limiting performance. Dataflow execution can improve locality and reduce redundant memory traffic, yet existing solutions are hardware-specific and incompatible with general-purpose SIMT programming. We present DACO, a Dataflow-Aware Compilation Optimization method that extends SIMT compilers to exploit dataflow opportunities automatically. DACO identifies three common patterns—intra-block dataflow, inter-block dataflow, and compute-memory dataflow—through static memory access analysis and generates optimized code with minimal developer effort. Experiments on real AI models show that DACO improves performance by up to 50.1% over the baseline SIMT compiler, highlighting its effectiveness and practicality.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

DACO: Unlocking Latent Dataflow Opportunities in Edge-Side SIMT Accelerators

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Han Zhao,
  • Yiying Xiang,
  • Yu Liu,
  • Xiaochun Ye,
  • Deze Zeng,
  • Jing Yang,
  • Weihao Cui,
  • Quan Chen,
  • Jingwen Leng,
  • Minyi Guo

摘要

Edge AI accelerators commonly use SIMT architectures, but AI kernels often suffer from high memory access overhead, limiting performance. Dataflow execution can improve locality and reduce redundant memory traffic, yet existing solutions are hardware-specific and incompatible with general-purpose SIMT programming. We present DACO, a Dataflow-Aware Compilation Optimization method that extends SIMT compilers to exploit dataflow opportunities automatically. DACO identifies three common patterns—intra-block dataflow, inter-block dataflow, and compute-memory dataflow—through static memory access analysis and generates optimized code with minimal developer effort. Experiments on real AI models show that DACO improves performance by up to 50.1% over the baseline SIMT compiler, highlighting its effectiveness and practicality.