RTLCurator: Label-Efficient Data Curation for RTL Generation

2026-07-31

AI Summary

Existing synthetic RTL datasets are large in scale but suffer from low correctness rates. Traditional functional testing filters often discard structurally complex yet syntactically valid samples, biasing training data toward simpler modules. This work proposes RTLCurator, which introduces, for the first time, a behavior-aware compatibility prior by modeling discrepancies between specifications and simulation-failed implementations. Combined with few-shot calibration using a small set of verified samples, RTLCurator constructs a high-quality subset that balances alignment, coverage, and structural diversity. Evaluations show that with only 10% of the data verified, the method retains 80% of high-quality samples and significantly outperforms full-data training, random selection, and pure simulation-based filtering on both CodeV and RTLCoder benchmarks.

Abstract

Training large language models (LLMs) to write register-transfer level (RTL) requires large corpora of paired specifications and code, and such data is scarce enough that most public corpora are now synthesized. Synthesis provides scale but not correctness, and in two widely used RTL datasets only 24.4% and 53.5% of pairs pass generated functional tests. This raises the question of how much of such a corpus to keep and which part of it. Correctness alone is a poor answer. A pair that misbehaves in one corner case still shows valid syntax and interface conventions, and complex sequential designs are both harder to generate and harder to validate, so filtering by correctness leaves a corpus of short and simple modules. Correctness is also hard to obtain, since behavior leaves little trace on the surface in RTL, and validating an entire corpus only sorts pairs into passed and failed. We present RTLCurator, which learns a behavior-aware compatibility prior by contrasting each specification with implementations that fail simulation, and calibrates it to a new corpus using a small number of validated pairs. It then constructs the retained subset by balancing alignment, representation coverage, and RTL structural richness. On CodeV and RTLCoder, keeping 80% of the corpus this way improves on training with the full corpus across all reported metrics while validating only 10% of the pool, whereas ranking by the score alone falls below random selection and filtering the whole pool by simulation does no better.

Problem

• RTL generation
• data curation
• label efficiency
• specification-code alignment
• behavioral correctness

Innovation

• data curation
• label-efficient learning
• RTL generation
• behavior-aware compatibility
• contrastive learning