Hot-FV

This repository contains the open source implementation of Hot-FV.

Envrionment

We have tested our work in Ubuntu-20.04 and Centos7. If you encounter some package missing in Python, simply install it.

How to Run

To execute Hot-FV, use the following commands:

WITHOUT_IMD=0 WITHOUT_ORD=0 BATCH=1 make b01
WITHOUT_IMD=0 WITHOUT_ORD=0 BATCH=1 make b06
WITHOUT_IMD=0 WITHOUT_ORD=0 BATCH=1 make b10
WITHOUT_IMD=0 WITHOUT_ORD=0 BATCH=1 make b11
WITHOUT_IMD=0 WITHOUT_ORD=0 BATCH=1 make b12
WITHOUT_IMD=0 WITHOUT_ORD=0 BATCH=1 make b14
WITHOUT_IMD=0 WITHOUT_ORD=0 BATCH=1 make or1200_Exception
WITHOUT_IMD=0 WITHOUT_ORD=0 BATCH=1 make or1200_ICache
WITHOUT_IMD=0 WITHOUT_ORD=0 BATCH=1 make or1200_DCache
WITHOUT_IMD=0 WITHOUT_ORD=0 BATCH=1 make or1200
WITHOUT_IMD=0 WITHOUT_ORD=0 BATCH=1 make RocketTile_Small
WITHOUT_IMD=0 WITHOUT_ORD=0 BATCH=1 make RocketTile_Tiny

Results will be saved in benchmark/[target]/result_[BATCH].

Folder Structure Explanation

Within the target directory of benchmark:

• cfg.pkl: Contains the parsed control flow graph of the design.
• idx.txt: Maps the corresponding id to the target branch id.

Within the exp_ folder in the target of benchmark:

• id: Contains the Pono result for each branch.
• last_state_value.txt: Contains the warm starting state for test generation for other branches.
• memory.txt: Contains memory results.
• result.txt: Shows branch coverage of our experiment (excluding reset related branches).
• unknown.txt and unreachable.txt: Indicate unknown and unreachable targets.
• score.txt: Logs the running process, including warm state selection.

Citation

Ziyue Zheng*, Zhiyuan Yan*, Xiangchen Meng, Guangyu Hu, Hongce Zhang, Yangdi Lyu, Hot-FV: A Semi-Formal Test Generation Framework for RTL Functional Coverage using Warm Starting States. ICCD 2025.

*: Same Contribution