Welcome to the Ace-TN Documentation

Ace-TN is a framework for simulating infinite projected entangled-pair states (iPEPS) with a focus on GPU acceleration. We follow a simple pythonic design philosophy with PyTorch tensors as the primary tensor-network building blocks, allowing users to quickly setup and manage iPEPS calculations on GPUs.

See our preprint at https://arxiv.org/abs/2503.13900.

Features

• Fast iPEPS ground-state calculations using GPUs.

• Flexible interface for custom model hamiltonian design.

• Modular and extensible design enabling quick integration of new iPEPS algorithms in a multi-GPU setting.

Code Example

The basic workflow of an iPEPS ground-state calculation with Ace-TN can be executed with

from acetn.ipeps import Ipeps
ipeps = Ipeps(config)
ipeps.evolve(dtau=0.01, steps=500)
measurements = ipeps.measure()

Please see the User Guide for more info on usage and getting started.

Citation

If you found Ace-TN useful for your research, please cite:

@misc{AceTN_2025,
      title={Ace-TN: GPU-Accelerated Corner-Transfer-Matrix Renormalization of Infinite Projected Entangled-Pair States},
      author={Addison D. S. Richards and Erik S. Sørensen},
      year={2025},
      eprint={2503.13900},
      archivePrefix={arXiv},
      primaryClass={cond-mat.str-el},
      url={https://arxiv.org/abs/2503.13900},
}