simuq 0.3.0

A Python package for quantum simulation with analog compilation

SimuQ

SimuQ is a framework for programming quantum simulations, compile them with analog pulses, and deploy the pulse schedules on real or simulated quantum devices.

Our project website includes use cases of SimuQ.

We illustrate our design and benefits in our arXiv paper.

Installation

We recommend Python 3.9 or greater since many optional dependencies have a minimum Python version requirement.

We encourage using pip for SimuQ installation. To install the core components of SimuQ, run the following command in shell:

pip install simuq

Multiple platforms are supported by SimuQ through different APIs and you may optionally install them based on your needs. You may install all optional dependencies, though not recommended, by

pip install "simuq[all]"

Note that the [braket] extra currently has package dependency conflicts with the [dwave] extra, so by installing the [all] extra we do not install the [dwave] extra. You need to manually install `[dwave]' if you want to use D-Wave backends.

Amazon Braket

SimuQ supports compilation to IonQ's trapped-ion devices and QuEra's neutral atom devices through Amazon Braket. Install the Amazon Braket provider and its dependencies by running

pip install "simuq[braket]"

If running on QPUs, make sure that your AWS account is onboarded to Amazon Braket, as per the instructions here (this isn't necessary for running on the local simulator).

IonQ Quantum Cloud

SimuQ supports compilation to IonQ's trapped-ion devices through IonQ Quantum Cloud. Install the IonQ provider and its dependencies by running

pip install "simuq[ionq]"

To run through IonQ Quantum Cloud, you must obtain an API key from IonQ. When creating IonQ providers in SimuQ, you must provide the API key either through a string or a file storing the key.

Qiskit

SimuQ supports compilation to IBM's superconducting devices through Qiskit and IBM Quantum. Install the Qiskit provider and its dependencies by running

pip install "simuq[qiskit]"

To run through IBM Quantum, you must obtain an API token from IBM.

D-Wave Leap

SimuQ supports compilation to D-Wave's superconducting devices through D-Wave Leap. Install the D-Wave provider and its dependencies by running

pip install "simuq[dwave]"

Note that this extra now has package dependency conflicts with the [braket] extra. Hence, users cannot simultaneously run SimuQ through D-Wave providers and Braket providers, and users need to choose which extra to install, if any. This is a known issue for the hardware providers and may need a period of time to be fixed.

The programming of Hamiltonian systems for D-Wave devices is also different. Due to the limited controllability of D-Wave's devices, currently the programming is limited to the final Hamiltonian of the evolution, and the annealing schedule can be directly passed to D-Wave providers in compilation.

QuTiP

You can simulate the dynamics of a SimuQ quantum system with QuTiP. Install the QuTiP provider and its dependencies by running

pip install "simuq[qutip]"

Installing from source

You can also install from source by cloning this repository and running a pip install command in the root directory of the repository:

git clone [e-mail]:PicksPeng/SimuQ.git
cd SimuQ
pip install .

If installing extras, remember to include quotes around the target (for example, pip install ".[dev]"). To install in editable mode, run pip install with the -e flag: pip install -e ".[dev]".

Examples

For examples of SimuQ usage, refer to the notebooks in the tutorials folder.

Project Structure

notebooks/: Example notebooks of running SimuQ and obtaining results.

notebooks/tutorials: Tutorial notebooks on usage of SimuQ.

notebooks/artifact_evaluation: AE notebooks to reproduce the data in our paper.

src/simuq/: The core compiler and language implementation of SimuQ.

src/simuq/aais/: AAISs of many backends in AAIS Specification Language.

src/simuq/backends/: The translation to API languages of different machine backends.

src/simuq/systems/: Hamiltonian systems implemented in Hamiltonian Modeling Language.

src/simuq/systems/benchmark/: A small benchmark of quantum Hamiltonian simulation as reported in our paper.

src/simuq/ionq/: The IonQ provider using IonQ Quantum Cloud.

src/simuq/braket/: The Braket provider using AWS Braket.

src/simuq/qiskit/: The Qiskit provider using IBM Quantum.

src/simuq/dwave/: The D-Wave provider using D-Wave Leap.

src/simuq/qutip/: The QuTiP provider using QuTiP.

Related projects

Please also check several projects related to SimuQ:

• QHDOPT: A software package for non-linear and non-convex continuous optimization using quantum backends. It employs SimuQ to program and deploy Quantum Hamiltonian Descent algorithm to find the minimum value of non-convex functions with box constraints.

Citations

If you use SimuQ in your work, please cite our paper.

@article{peng2024simuq,
	author = {Peng, Yuxiang and Young, Jacob and Liu, Pengyu and Wu, Xiaodi},
	title = {SimuQ: A Framework for Programming Quantum Hamiltonian Simulation with Analog Compilation},
	year = {2024},
	issue_date = {January 2024},
	publisher = {Association for Computing Machinery},
	address = {New York, NY, USA},
	volume = {8},
	number = {POPL},
	url = {https://doi.org/10.1145/3632923},
	doi = {10.1145/3632923},
	month = {jan},
	articleno = {81},
	numpages = {31},
}