Sequential Logic Datasets with Designs

Collection of datasets and design files for sequential digital circuits, used to train and evaluate neural networks that duplicate circuit functionality.

Table of contents

• Overview
• Quickstart
• Repository contents
• Dataset format
• How to generate new datasets
• Usage with main project
• License
• Contact / Acknowledgements

Overview

This repository contains sequential logic designs (VHDL and Quartus schematic files), testbenches, and corresponding datasets of input/output traces. These datasets are used by the companion project Digital-IC-Functionality-Duplication-Using-NN to train neural-network models that approximate digital circuit behavior.

Quickstart

1. Clone the repository:

git clone https://github.com/Anjanamb/Sequential-Logic-Datasets-with-Designs.git
cd Sequential-Logic-Datasets-with-Designs

2. Browse the available circuit designs and their corresponding datasets.

3. Copy the required dataset .txt files into the data/ folder of the main project for training.

Repository contents

Sequential-Logic-Datasets-with-Designs/
├─ Designs/          # VHDL designs, Quartus schematic (.bdf), and testbenches (.vhd)
├─ Datasets/         # Generated input/output/state traces in text format
├─ Scripts/          # Simulation and dataset-preparation scripts
└─ README.md

• Designs/ – contains source circuit designs (flip-flops, counters, shift registers, etc.).
• Datasets/ – plain-text traces derived from testbenches (aligned inputs and outputs).
• Scripts/ – helper scripts to run simulations and extract traces.

Dataset format

Each dataset file (.txt) contains aligned columns of inputs, states, and outputs observed from the circuit simulation.

Typical format:

clk in1 in2 state out1 out2
0   0   1   00    0    1
1   1   0   01    1    0
...

• clk: clock cycle
• inX: circuit inputs
• state: encoded internal state (binary string or integer)
• outX: observed outputs

How to generate new datasets

1. Modify or create a new VHDL/Quartus design in Designs/.
2. Write a testbench (.vhd) to simulate input patterns.
3. Run simulation in ModelSim/Quartus and export results.
4. Convert the simulation output to a plain .txt trace (using scripts in Scripts/).
5. Place the generated dataset in Datasets/.

Usage with main project

To use these datasets with the neural-network training project:

1. Generate or select a dataset from Datasets/.
2. Copy or symlink it into the data/ folder of Digital-IC-Functionality-Duplication-Using-NN.
3. Pass the dataset path to the training script (see main project README).

License

MIT License – feel free to use, modify, and share. Please credit the repository if used in research or academic work.

Contact / Acknowledgements

Author: Anjana Bandara (GitHub: Anjanamb)
Contributors: Ayesh-Rajakaruna, sahannt98