AI

AgRefactor: Self-Evolving Agentic Workflow for HLS Compatibility and Performance

Researchers have developed a new AI system called AgRefactor that can automatically convert real-world software into code compatible with High-Level Synthesis (HLS) systems. This is done through a multi-agent workflow that incorporates a self-evolving memory system and integrates automated refactoring tools to improve efficiency and scalability. According to the authors, AgRefactor outperforms or matches existing state-of-the-art methods on 9 out of 11 challenging real-world
Researchers have developed a new AI system called AgRefactor that can automatically convert real-world software into code compatible with High-Level Synthesis (HLS) systems. This is done through a multi-agent workflow that incorporates a self-evolving memory system and integrates automated refactoring tools to improve efficiency and scalability. According to the authors, AgRefactor outperforms or matches existing state-of-the-art methods on 9 out of 11 challenging real-world benchmarks, with some tests showing significant speedup over previous approaches. --- Why it matters: This matters because it addresses a long-standing challenge in converting software into HLS-compatible code, which is crucial for rapid development and deployment of hardware systems. The ability to automate this process efficiently can significantly impact the productivity of engineers working on such projects. Source: https://arxiv.org/abs/2606.30949

This article was originally published at: https://arxiv.org/abs/2606.30949