DAPO: Optimizing High-Level Synthesis with AI-Driven Pass Ordering
Published:Dec 12, 2025 07:35
•1 min read
•ArXiv
Analysis
This research explores a novel application of AI in optimizing the pass ordering within high-level synthesis (HLS), potentially leading to significant performance improvements in hardware design. The use of graph contrastive and reinforcement learning techniques suggests a sophisticated approach to addressing a complex optimization problem in the field.
Key Takeaways
- •Applies AI, specifically graph contrastive learning and reinforcement learning, to optimize pass ordering in High-Level Synthesis.
- •Aims to improve performance in hardware design through intelligent pass scheduling.
- •Targets a significant and complex optimization problem within the HLS workflow.
Reference
“DAPO employs Graph Contrastive and Reinforcement Learning.”