The Cream Rises to the Top: Efficient Reranking Method for Verilog Code Generation

2025/09/24 by Yang, Guang, Zheng, Wei, Chen, Xiang +3 · 1 citation
#Artificial Intelligence (cs.AI) #FOS: Computer and information sciences #Hardware Architecture (cs.AR) #Software Engineering (cs.SE)

paper · doi:10.48550/arxiv.2509.20215

Abstract

LLMs face significant challenges in Verilog generation due to limited domain-specific knowledge. While sampling techniques improve pass@k metrics, hardware engineers need one trustworthy solution rather than uncertain candidates. To bridge this gap, we formulate it as a semantic alignment problem between requirements and Verilog implementations, and propose VCD-RNK, a discriminator model tailored for efficient Verilog code reranking. Specifically, VCD-RNKincorporates Verilog-specific reasoning by distilling expert knowledge across three dimensions: code semantic analysis, test case generation, and functional correctness assessment. By explicitly simulating the above reasoning processes during inference, VCD-RNK effectively avoids computationally intensive test execution in existing methods.

Citations

• Seed-Coder: Let the Code Model Curate Data for Itself
• VeriCoder: Enhancing LLM-Based RTL Code Generation through Functional Correctness Validation
• DeepRTL: Bridging Verilog Understanding and Generation with a Unified Representation Model
• VerilogCoder: Autonomous Verilog Coding Agents with Graph-based Planning and Abstract Syntax Tree (AST)-based Waveform Tracing Tool
• Continual Learning of Large Language Models: A Comprehensive Survey
• Towards Cross-Domain Continual Learning
• Chain-of-Thought in Neural Code Generation: From and For Lightweight Language Models
• VerilogEval: Evaluating Large Language Models for Verilog Code Generation
• Chip-Chat: Challenges and Opportunities in Conversational Hardware Design
• CodeT: Code Generation with Generated Tests
• LoRA: Low-Rank Adaptation of Large Language Models

Cited by

• Large Language Model for Verilog Code Generation: Literature Review and the Road Ahead

Related

• VFocus: Better Verilog Generation from Large Language Model via Focused Reasoning
• VRank: Enhancing Verilog Code Generation from Large Language Models via Self-Consistency
• QiMeng-CodeV-R1: Reasoning-Enhanced Verilog Generation
• Insights from Verification: Training a Verilog Generation LLM with Reinforcement Learning with Testbench Feedback
• VerilogMonkey: Exploring Parallel Scaling for Automated Verilog Code Generation with LLMs