QuCLEAR

Artifact for the paper: QuCLEAR: Clifford Extraction and Absorption for Quantum Circuit Optimization

---

Installation

We recommend creating an Anaconda virtual environment before proceeding with the installation.
Anaconda Installation Guide

1. Clone the GitHub Repository

First, clone the QuCLEAR repository and navigate to its directory:

git clone https://github.com/revilooliver/QuCLEAR.git
cd QuCLEAR

---

2. Install QuCLEAR

If you only wish to use QuCLEAR (without running the artifact evaluation against other tools), you can install the required dependencies with:

pip install -r requirements.txt

You may skip Sections 3 and 4 and instead refer to the examples in the tutorial folder.

---

3. Environment Setup for Artifact Evaluation

For artifact evaluation, QuCLEAR is compared with several tools: Rustiq, Qiskit, Paulihedral, and pytket. To simplify installation and environment management, we provide an automated script, install_all.sh, located in the artifact_evaluation folder:

cd artifact_evaluation
./install_all.sh

This script creates three Anaconda virtual environments:

• QuCLEAR_env: Contains QuCLEAR, Qiskit, and pytket.
• PH_env: Contains Paulihedral.
• Tetris_env: Contains Tetris.

Install Rust and Rustiq in the QuCLEAR_env

After setting up the environments, you will need to manually install Rust and Rustiq in the QuCLEAR_env.

1. Activate the QuCLEAR environment:

   conda activate QuCLEAR_env

2. Install Rust:
   Follow the installation instructions provided on the official Rust website:
   Rust Installation Guide

   After installation, source your .bashrc file to enable Rust in your current session:

   source ~/.bashrc

   If you are working in a Conda environment, you may need to reactivate the environment to ensure the paths are properly configured:

   conda deactivate
   conda activate QuCLEAR_env

3. Install Rustiq:
   Clone and install Rustiq from its GitHub repository:
   Rustiq Repository

Once these steps are complete, all required dependencies for artifact evaluation will be ready.

4. Test Installation

We provide a simple test script to validate the installation. It will compile the circuits for Hamiltonian simulation:

./test_experiments_fast.sh

We also provide a simple test script to validate the experiments for mapping to hardware:

./test_mapping_experiments_fast.sh

---

Validating the Experiments

The experimental data can be validated by running the full experiment scripts:

cd artifact_evaluation
./test_experiments_full.sh
./test_mapping_experiments_full.sh

Results

The compilation results are saved as JSON files in the experiments/results_fullyconnected folder.
The hardware mapping results are also saved as JSON files in the experiments/results_ibm and experiments/results_google folder

For example:

• test_quclear_LiH.json contains results for QuCLEAR, Qiskit, and Rustiq for the LiH benchmark.

Data Explanation

• our_method: Results using QuCLEAR.
• combined_method: Results using QuCLEAR combined with Qiskit's optimization (reported in the paper).

---

Generating Table Data

To generate the experimental tables in our paper, use the provided Jupyter notebooks:

1. generate_table3.ipynb: Generates data for Table 3.
2. generate_table4.ipynb: Generates data for Table 4.
3. generate_figure9.ipynb: Generates data for Figure 9.
4. generate_figure10.ipynb: Generates data for Figure 10.
5. generate_figure11_google.ipynb and generate_figure11_ibm.ipynb: Generates data for Figure 11.

These notebooks process results stored in the experiments/results_fullyconnected folder. The notebooks for figure11 process results in the experiments/results_ibm and experiments/results_google folders.

---

Tutorials

We provide two example tutorials in the tutorial folder.

1. Optimizing Circuits and Calculating Expectation Values

Run the VQE_observables.ipynb notebook to learn how to:

• Optimize quantum circuits.
• Calculate expectation values in variational quantum algorithms (e.g., VQE).

2. Absorbing CNOT Networks in QAOA

Run the QAOA_probabilities.ipynb notebook to see examples of:

• Absorbing CNOT networks.
• Optimizing circuits for QAOA.

License

This project is licensed under the Apache License 2.0