This project is a quantum circuit verification and simulation tool using TDDs[1] in conjunction with the contraction planning tool Cotengra[2]. The project uses parameterised circuits provided by MQT.Bench[3]. This project can only be run on Linux (Ubuntu 20.04) and using Python 3.9.
To ensure that the required python libraries are installed, run the following command from the root of the folder:
pip install -r requirements.txt
If the external repositories are not properly collected, run the following command:
git submodule update --init --recursive
To perform any of the experiments seen in the associated article, the main experiment file must be targeted from the root of the folder.
python3 src/experiment [arguments]
All arguments for the experiments can be seen below.
An example of an experiment may be running the DJ algorithm with qubits 64, 128, and 256 using Cotengra's RGreedy heuristic
python3 src/experiment --exp_name DJ_RGreedy_Benchmark --exp_type formal --setup standard --algorithms dj --level 0,2 --qubits 64 128 256
| Name | Values | Description |
|---|---|---|
| --exp_name | any string | The name of the experiment and folder where data is saved |
| --exp_type | formal, simulation | Determines the setup for the experiment |
| -counter | Flag for whether a counter example when inequality is determined should be found | |
| -ts | Flag for whether to use time stamp for folder name | |
| --setup | standard, pure_qcec | Denotes the type of method to test (standard is article method) |
| --algorithms | ghz, dj, graphstate, twolocalrandom, qftentangled, qpeexact, su2random, wstate, realamprandom | The circuit algorithm used for the setup |
| --levels | any two-tuple of 2 ints in [0,3], e.g. 0,2 | The abstraction levels for each of the two circuit |
| --qubit_range | any two-tuple of 2 ints | The start and stop of number of qubits |
| --qubits | list of ints | The exact number of qubits to test for |
| --gate_deletions | int | The number of random gates to delete. Must be greater than or equal to zero. Cannot be the entire circuit. |
| --reps | int | Number of repetitions |
| -split | Flag for splitting the multi-qubit gates, e.g. CNOT | |
| -subnets | Flag for using subnetworks | |
| --max_cont_time | int | Max time allowed for contraction, -1 for no limit |
| --method | cotengra, linear | The overall method used for planning |
| --path_method | rgreedy, betweenness, walktrap, linear, proportional | The heuristic used for planning. Depends on the method selected |
| --minimize | flops, size, combo | The minimization goal for cotengra when finding plans |
| --max_repeats | int | The number of repeats plans found by cotengra |
| --max_plan_time | int | Max allowed time for planning (soft constraint). |
[1] Xin Hong, Xiangzhen Zhou, Sanjiang Li, Yuan Feng, and Mingsheng Ying (2021). A tensor network based decision diagram for representation of quantum circuits.
[2] Johnnie Gray and Stefanos Kourtis (2021). Hyper-optimized tensor network contraction. Quantum, 5:410, mar 2021.
[3] Nils Quetschlich, Lukas Burgholzer, and Robert Wille (2023). MQT Bench: Benchmarking software and design automation tools for quantum computing. Quantum, 2023.