# Agent Factories for High Level Synthesis: How Far Can General-Purpose Coding Agents Go in Hardware Optimization?

arXiv: https://arxiv.org/abs/2603.25719 · HTML: https://mlsanity.com/paper/2603.25719/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in Software engineering coding agents, one of 16 directions in LLM agents and tool use.
- Builds on 18 of its 35 references that are in the map.
- First posted on arXiv 26 Mar 2026.

## Placed in

- LLM agents and tool use / [Software engineering coding agents](https://mlsanity.com/agents/software-engineering-coding-agents/)

## Builds on (most cited first, 18 of 18)

- [MetaGPT: Meta Programming for A Multi-Agent Collaborative Framework](https://mlsanity.com/paper/2308.00352/) (2023)
- [SWE-agent: Agent-Computer Interfaces Enable Automated Software Engineering](https://mlsanity.com/paper/2405.15793/) (2024)
- [ChatDev: Communicative Agents for Software Development](https://mlsanity.com/paper/2307.07924/) (2023)
- [IronMan: GNN-assisted Design Space Exploration in High-Level Synthesis via Reinforcement Learning](https://mlsanity.com/paper/2102.08138/) (2021)
- [AutoDSE: Enabling Software Programmers to Design Efficient FPGA Accelerators](https://mlsanity.com/paper/2009.14381/) (2020)
- [HLSPilot: LLM-based High-Level Synthesis](https://mlsanity.com/paper/2408.06810/) (2024)
- [Rollout Roulette: A Probabilistic Inference Approach to Inference-Time Scaling of LLMs using Particle-Based Monte Carlo Methods](https://mlsanity.com/paper/2502.01618/) (2025)
- [Automated C/C++ Program Repair for High-Level Synthesis via Large Language Models](https://mlsanity.com/paper/2407.03889/) (2024)
- [C2HLSC: Leveraging Large Language Models to Bridge the Software-to-Hardware Design Gap](https://mlsanity.com/paper/2412.00214/) (2024)
- [Confucius Code Agent: Scalable Agent Scaffolding for Real-World Codebases](https://mlsanity.com/paper/2512.10398/) (2025)
- [Cross-Modality Program Representation Learning for Electronic Design Automation with High-Level Synthesis](https://mlsanity.com/paper/2406.09606/) (2024)
- [LLM-DSE: Searching Accelerator Parameters with LLM Agents](https://mlsanity.com/paper/2505.12188/) (2025)
- [Can Reasoning Models Reason about Hardware? An Agentic HLS Perspective](https://mlsanity.com/paper/2503.12721/) (2025)
- [LIFT: LLM-Based Pragma Insertion for HLS via GNN Supervised Fine-Tuning](https://mlsanity.com/paper/2504.21187/) (2025)
- [iDSE: Navigating Design Space Exploration in High-Level Synthesis Using LLMs](https://mlsanity.com/paper/2505.22086/) (2025)
- [Automatic Hardware Pragma Insertion in High-Level Synthesis: A Non-Linear Programming Approach](https://mlsanity.com/paper/2405.12304/) (2024)
- [Agentic-HLS: An agentic reasoning based high-level synthesis system using large language models (AI for EDA workshop 2024)](https://mlsanity.com/paper/2412.01604/) (2024)
- [AutoHLS: Learning to Accelerate Design Space Exploration for HLS Designs](https://mlsanity.com/paper/2403.10686/) (2024)

## Built on by (most cited first, 1 of 1)

- [HLSmith: An Expert-Guided Agentic Framework for C/C++-to-HLS Translation](https://mlsanity.com/paper/2608.06791/) (2026)

## Abstract (arXiv)

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 assemble globally promising configurations under an area constraint. In Stage~2, it launches N expert agents over the top ILP solutions, each exploring cross-function optimizations such as pragma recombination, loop fusion, and memory restructuring that are not captured by sub-kernel decomposition. We evaluate the approach on 12 kernels from HLS-Eval and Rodinia-HLS using Claude Code (Opus~4.5/4.6) with AMD Vitis HLS. Scaling from 1 to 10 agents yields a mean 8.27× speedup over baseline, with larger gains on harder benchmarks: streamcluster exceeds 20× and kmeans reaches approximately 10×. Across benchmarks, agents consistently rediscover known hardware optimization patterns without domain-specific training, and the best designs often do not originate from top-ranked ILP candidates, indicating that global optimization exposes improvements missed by sub-kernel search. These results establish agent scaling as a practical and effective axis for HLS optimization.