This repository is a Python module of Pauli decomposition algorithms [1-4] implemented with C++.
Our C++ implementation is mostly based on the Python implementation by Jones, except for the iterative algorithm by Hantzko et al. [2] since it is not included in the Python implementation.
The C++ library Boost needs to be installed.
The mprof command is also required to run the memory benchmark.
It is recommended to prepare and use a new virtual Python environment.
You need to run pip3 install -r requirements.txt to set up the module. You may need to change the version of the Boost.Python and Boost.Numpy in setup.py.
You can test the module with python3 -m pytest tests.
The benchmarks directory contains a set of files for benchmarking the algorithms.
You can benchmark the runtime of the algorithms by the following commands. You can terminate the second command anytime.
cd benchmarks
python3 measure.py output_file.txt
python3 draw.py output_file.txtYou can check the memory consumption of the algorithms by python memory.py [# of qubits].
[1] Hiroki Hamaguchi, Kou Hamada, Nobuyuki Yoshioka. “Handbook for Efficiently Quantifying Robustness of Magic” (2023) arXiv:2311.01362.
[2] Lukas Hantzko, Lennart Binkowski, and Sabhyata Gupta. “Tensorized Pauli decomposition algorithm” (2023) arXiv:2310.13421.
[3] Tyson Jones. “Decomposing dense matrices into dense Pauli tensors” (2024). arxiv:2401.16378.
[4] Sebastián Vidal Romero and Juan Santos-Suárez. “PauliComposer: Compute tensor products of Pauli matrices efficiently”. Quantum Information Processing 22, 449 (2023).
Copyright (c) 2024 Nobuyuki Yoshioka
This project is licensed under the MIT License - see the LICENSE file for details.