arXiv subjects

Dongsheng Zuo

Publications and source records attributed to Dongsheng Zuo.

6 recordsLinked to original sources

arXiv2026

HINT: Toward an Executable Hardware-Intent Representation Layer for LLM-Driven RTL Generation

Generating implementation-quality RTL with large language models (LLMs) remains difficult because direct generation must resolve microarchitecture while simultaneously producing and debugging low-level code. We present HINT, an executable hardware-intent intermediate representation layer between behavioral specifications or executable oracles and RTL. HINT makes RTL-relevant microarchitecture explicit, supports pre-RTL checking, and supplies explicit RTL-lowering obligations. We evaluate HINT using both a minimal single-agent flow and a full staged workflow. Across seven operator cases, the HINT-mediated route, with no post-synthesis QoR refinement, produces contract-compliant synthesizable RTL on 7/7 cases; Direct C2RTL and C2HLSC apply to five cases and succeed on 5/5 and 1/5, respectively. Under matched Design Compiler synthesis, HINT reduces area by 5.0\%--26.2\% relative to five manual RTL implementations and by 8.9\%--86.1\% relative to five accepted Direct C2RTL results. RealBench AES and SDC, together with a Vortex VPU synthesizing to 561.67k~$\mu\mathrm{m}^2$, further demonstrate specification-driven, protocol-rich, memory-rich, and hierarchical designs. In the controlled operator study, the HINT-mediated route shows better observed convergence and avoids the severe implementation-quality degradation seen in several direct-generation results.

cs.AR↗

arXiv2026

Retrieve, Schedule, Reflect: LLM Agents for Chip QoR Optimization

Modern chip design requires multi-objective optimization of timing, power, and area under stringent time-to-market constraints. Although powerful optimization algorithms are integrated into EDA tools, achieving high QoR hinges on effective long-horizon scheduling, which relies heavily on manual expert intervention. To address this issue and automate chip design, we propose an agentic LLM framework that schedules chip optimizations through direct interaction with EDA tools. The agent is grounded in natural language expertise expressed as a search tree through retrieval-augmented generation (RAG). We further improve scheduling quality with Pareto-driven QoR feedback through language reflection. Experimental results show that, compared with black-box search methods such as reinforcement learning, our framework achieves 10% greater timing improvement while consuming less power and area, with more than 4x speedup. The post-optimization QoR is also comparable to that achieved by human experts. Finally, the agent supports customized tasks expressed in natural language, enabling preferential QoR trade-offs. The code and chip design data will be publicly available at https://github.com/YiKangOY/Open-LLM-ECO.

cs.AR↗

arXiv2025

PrefixAgent: An LLM-Powered Design Framework for Efficient Prefix Adder Optimization

Prefix adders are fundamental arithmetic circuits, but their design space grows exponentially with bit-width, posing significant optimization challenges. Previous works face limitations in performance, generalization, and scalability. To address these challenges, we propose PrefixAgent, an LLM-powered framework that enables efficient prefix adder optimization. Firstly, we reformulate the problem into two subtasks, namely backbone synthesis and structure refinement, which effectively reduce the search space. The LLM performs these two phases of optimization by invoking tools through function calls, enabling it to iteratively construct the backbone and refine local structures based on reasoning and EDA feedback. Secondly, this new design perspective allows us to systematically collect large-scale, high-quality supervision data. We leverage the rewriting and equality saturation capabilities of e-graphs to comprehensively explore the prefix adder solution space. In addition, the explainability of e-graphs enables us to extract fine-grained rewrite trajectories, which serve as interpretable and effective training data. PrefixAgent is then fine-tuned on this dataset, significantly enhancing its optimization ability and reasoning generalization. Experimental results show that PrefixAgent synthesizes prefix adders with smaller areas than baseline methods in nearly all configurations, with the advantage growing at larger bit-widths, and it remains effective under a commercial EDA flow.

cs.AR↗

arXiv2024

RL-MUL 2.0: Multiplier Design Optimization with Parallel Deep Reinforcement Learning and Space Reduction

Multiplication is a fundamental operation in many applications, and multipliers are widely adopted in various circuits. However, optimizing multipliers is challenging due to the extensive design space. In this paper, we propose a multiplier design optimization framework based on reinforcement learning. We utilize matrix and tensor representations for the compressor tree of a multiplier, enabling seamless integration of convolutional neural networks as the agent network. The agent optimizes the multiplier structure using a Pareto-driven reward customized to balance area and delay. Furthermore, we enhance the original framework with parallel reinforcement learning and design space pruning techniques and extend its capability to optimize fused multiply-accumulate (MAC) designs. Experiments conducted on different bit widths of multipliers demonstrate that multipliers produced by our approach outperform all baseline designs in terms of area, power, and delay. The performance gain is further validated by comparing the area, power, and delay of processing element arrays using multipliers from our approach and baseline approaches.

cs.AR↗

arXiv2024

UFO-MAC: A Unified Framework for Optimization of High-Performance Multipliers and Multiply-Accumulators

Multipliers and multiply-accumulators (MACs) are critical arithmetic circuit components in the modern era. As essential components of AI accelerators, they significantly influence the area and performance of compute-intensive circuits. This paper presents UFO-MAC, a unified framework for the optimization of multipliers and MACs. Specifically, UFO-MAC employs an optimal compressor tree structure and utilizes integer linear programming (ILP) to refine the stage assignment and interconnection of the compressors. Additionally, it explicitly exploits the non-uniform arrival time profile of the carry propagate adder (CPA) within multipliers to achieve targeted optimization. Moreover, the framework also supports the optimization of fused MAC architectures. Experimental results demonstrate that multipliers and MACs optimized by UFO-MAC Pareto-dominate state-of-the-art baselines and commercial IP libraries. The performance gain of UFO-MAC is further validated through the implementation of multipliers and MACs within functional modules, underlining its efficacy in real scenarios.

cs.AR↗

arXiv2024

E-Syn: E-Graph Rewriting with Technology-Aware Cost Functions for Logic Synthesis

Logic synthesis plays a crucial role in the digital design flow. It has a decisive influence on the final Quality of Results (QoR) of the circuit implementations. However, existing multi-level logic optimization algorithms often employ greedy approaches with a series of local optimization steps. Each step breaks the circuit into small pieces (e.g., k-feasible cuts) and applies incremental changes to individual pieces separately. These local optimization steps could limit the exploration space and may miss opportunities for significant improvements. To address the limitation, this paper proposes using e-graph in logic synthesis. The new workflow, named Esyn, makes use of the well-established e-graph infrastructure to efficiently perform logic rewriting. It explores a diverse set of equivalent Boolean representations while allowing technology-aware cost functions to better support delay-oriented and area-oriented logic synthesis. Experiments over a wide range of benchmark designs show our proposed logic optimization approach reaches a wider design space compared to the commonly used AIG-based logic synthesis flow. It achieves on average 15.29% delay saving in delay-oriented synthesis and 6.42% area saving for area-oriented synthesis.

cs.AR↗