OpenABC-D

该数据集是一个大规模的标注数据集，它通过结合开源设计与一款领先的开源逻辑综合工具合成而成，包含了87万个And-Inverter-Graphs（AIGs），这些AIGs源自1500次综合运行。数据集的生成流程包括三个阶段：寄存器传输级（RTL）综合、图级别处理以及为机器学习进行的预处理。这一流程不仅能够实现自动化综合脚本生成，还允许用户定义综合脚本。该数据集的规模达到了87万个AIGs，来自1500次综合运行，其任务是利用机器学习进行逻辑综合优化。

This is a large-scale labeled dataset synthesized by combining open-source designs with a leading open-source logic synthesis tool. It contains 870,000 And-Inverter-Graphs (AIGs) derived from 1500 synthesis runs. The dataset generation workflow comprises three stages: Register-Transfer Level (RTL) synthesis, graph-level processing, and preprocessing tailored for machine learning. This workflow supports automated synthesis script generation and allows users to define custom synthesis scripts. With a corpus of 870,000 AIGs from 1500 synthesis runs, this dataset is designed for logic synthesis optimization using machine learning techniques.