This repository hosts an (BSD 3-clause) open source framework for running contraction sequence algorithms for tensor network simulations. These algorithms come from two primary sources: The tensor network literature and the structural graph theory and algorithms literature. In this framework we provide pre- and post-processing routines for wrangling data, an algorithm dispatcher for easily running batch experiments, a command-line interface for computing a contraction sequence on a given network with a given algorithm, and a containerized method for running individual algorithms.
For notes on running our Dockerized treewidth solvers, see TREEWIDTH.md.
For detailed notes on installing and using ConSequences
, see the README in the consequences
subfolder.
For the experimental supplement to "Benchmarking treewidth as a practical component of tensor-network--based simulation", see the README in the paper
subfolder.
A Zenodo DOI will be provided once all code is finalized. In the mean time, please cite the corresponding paper on arXiv:
@article{dumitrescu2018benchmarking,
author = "Dumitrescu, Eugene F. and Fisher, Allison L. and
Goodrich, Timothy D. and Humble, Travis S. and Sullivan,
Blair D. and Wright, Andrew L.",
title = "{Benchmarking treewidth as a practical component of
tensor-network--based quantum simulation}",
year = "2018",
eprint = "1807.04599",
archivePrefix = "arXiv",
primaryClass = "cs.DS",
SLACcitation = "%%CITATION = ARXIV:1807.04599;%%"
}
Having problems installing or running code? Want to discuss codebase extensions or additional use cases? Contact us by creating a GitHub issue or by emailing Timothy at tdgoodri@ncsu.edu.