QFAST is no longer being actively maintained. The successor to qfast is BQSKit.
BQSKit combines several quantum synthesis projects, including the main synthesis algorithm implemented by qfast, qsearch, and a circuit partitioning and reoptimization project. BQSKit is being actively developed and with new features and bug fixes, including updates based on feedback from users.
BQSKit does have a different API, so moving an existing project to BQSKit may take some effort. In the meantime, qfast isn't going anywhere, but is not longer actively maintained, will not receive any new features, and may not receive bugfix or compatability updates.
QFAST is a quantum synthesis tool designed to produce short circuits and to scale well in practice. QFAST uses a mathematical model of circuits encoding both gate placement and function. This is packaged together with a hierarchical stochastic gradient descent formulation that combines “coarse-grained” fast optimization during circuit structure search with a better, but slower, stage only in the final circuit refinement.
The best way to install this python package is with pip.
pip install qfast
QFAST can be used to convert a quantum operation specified by a unitary matrix into a circuit given by openqasm code. There is a command-line interface provided with qfast that can be accessed by python -m qfast. This can be used to synthesize a matrix.
python -m qfast input.unitary output.qasm
Here the input.unitary file is a NumPy matrix saved with np.savetxt, the qasm output will be saved in the output.qasm file and the KAK native tool will be used. The command-line help option python -m qfast -h can be used for further information.
QFAST can also be used as a library, an example is included.
Native tools are necessary for QFAST to perform instantiation. During decomposition, the input unitary matrix is hierarchically broken into many smaller unitaries. At some level in the hierarchy, QFAST switches to instantiation, which uses a native synthesis tool to convert the small unitaries into native gates.
Included with this python package is the QSearch native tool. Here are some others:
- qfast-qiskit: Several qiskit tools
- qfast-uq: A UniversalQCompiler native tool (deprecated)
- qfast-qs: A QSearch native tool (Now default)
Younis, Ed, et al. "QFAST: Quantum Synthesis Using a Hierarchical Continuous Circuit Space." arXiv preprint arXiv:2003.04462 (2020).
Quantum Fast Approximate Synthesis Tool (QFAST) Copyright (c) 2020, The Regents of the University of California, through Lawrence Berkeley National Laboratory (subject to receipt of any required approvals from the U.S. Dept. of Energy). All rights reserved.
If you have questions about your rights to use or distribute this software, please contact Berkeley Lab's Intellectual Property Office at IPO@lbl.gov.
NOTICE. This Software was developed under funding from the U.S. Department of Energy and the U.S. Government consequently retains certain rights. As such, the U.S. Government has been granted for itself and others acting on its behalf a paid-up, nonexclusive, irrevocable, worldwide license in the Software to reproduce, distribute copies to the public, prepare derivative works, and perform publicly and display publicly, and to permit others to do so.