Demonstrates that general-purpose coding agents can achieve 20x speedups in hardware design optimization without domain-specific training.
March 27, 2026
Original Paper
Agent Factories for High Level Synthesis: How Far Can General-Purpose Coding Agents Go in Hardware Optimization?
arXiv · 2603.25719
The Takeaway
The two-stage agent factory approach rediscovers expert-level hardware optimization patterns (like loop fusion and memory restructuring) autonomously. This suggests that scaling agent-based search is a practical path to automating complex High-Level Synthesis (HLS) tasks.
From the abstract
We present an empirical study of how far general-purpose coding agents -- without hardware-specific training -- can optimize hardware designs from high-level algorithmic specifications. We introduce an agent factory, a two-stage pipeline that constructs and coordinates multiple autonomous optimization agents.In Stage~1, the pipeline decomposes a design into sub-kernels, independently optimizes each using pragma and code-level transformations, and formulates an Integer Linear Program (ILP) to ass