quantum-pecos 0.8.0.dev8

PECOS is a library for the evaluation, study, and design of quantum error correction protocols.

Performance Estimator of Codes On Surfaces (PECOS) is a library/framework dedicated to the study, development, and evaluation of quantum error-correction protocols. It also offers tools for the study and evaluation of hybrid quantum/classical compute execution models.

Initially conceived and developed in 2014 to verify lattice-surgery procedures presented in arXiv:1407.5103 and released publicly in 2018, PECOS filled the gap in the QEC/QC tools available at that time. Over the years, it has grown into a framework for studying general QECCs and hybrid computation.

Features

• Quantum Error-Correction Tools: Advanced tools for studying quantum error-correction protocols and error models.
• Hybrid Quantum/Classical Execution: Evaluate advanced hybrid compute models, including support for classical compute, calls to Wasm VMs, conditional branching, and more.
• Fast Simulation: Leverages a fast stabilizer simulation algorithm.
• Multi-language extensions: Core functionalities implemented via Rust for performance and safety. Additional add-ons and extension support in C/C++ via Cython.
• LLVM IR Support: Execute LLVM Intermediate Representation programs for hybrid quantum/classical computing. LLVM support is optional - PECOS can be built without LLVM by using --no-default-features when building the Rust crates. When LLVM is enabled (default), requires LLVM version 14.

Getting Started

Explore the capabilities of PECOS by delving into the documentation.

Repository Structure

PECOS now consists of multiple interconnected components:

• /python/: Contains Python packages
  • /python/quantum-pecos/: Main Python package (imports as pecos)
  • /python/pecos-rslib/: Python package with Rust extensions that utilize the pecos crate
• /crates/: Contains Rust crates
  • /crates/pecos/: Main Rust crate that collects the functionality of the other crates into one library
  • /crates/pecos-core/: Core Rust functionalities
  • /crates/pecos-qsims/: A collection of quantum simulators
  • /crates/pecos-qec/: Rust code for analyzing and exploring quantum error correction (QEC)
  • /crates/pecos-qasm/: Implementation of QASM parsing and execution
  • /crates/pecos-llvm-runtime/: Implementation of LLVM IR execution for hybrid quantum-classical programs
  • /crates/pecos-engines/: Quantum and classical engines for simulations
  • /crates/pecos/: Main PECOS library (includes CLI with cli feature)
  • /crates/pecos-build/: Developer tools CLI (LLVM setup, dependency management)
  • /crates/pecos-python/: Rust code for Python extensions
  • /crates/benchmarks/: A collection of benchmarks to test the performance of the crates
• /julia/: Contains Julia packages (experimental)
  • /julia/PECOS.jl/: Main Julia package
  • /julia/pecos-julia-ffi/: Rust FFI library for Julia bindings

Quantum Error Correction Decoders

PECOS includes LDPC (Low-Density Parity-Check) quantum error correction decoders as optional components. See DECODERS.md for detailed information about:

• LDPC decoder algorithms and variants
• How to build and use decoders
• Performance considerations
• Architecture and development guide

You may find most of these crates in crates.io if you wish to utilize only a part of PECOS, e.g., the simulators.

Versioning

We follow semantic versioning principles. However, before version 1.0.0, the MAJOR.MINOR.BUG format sees the roles of MAJOR and MINOR shifted down a step. This means potential breaking changes might occur between MINOR increments, such as moving from versions 0.1.0 to 0.2.0.

All Python packages and all Rust crates will have the same version amongst their respective languages; however, Python and Rust versioning will differ.

Latest Development

Stay updated with the latest developments on the PECOS Development branch.

Installation

Python Package

To install the main Python package for general usage:

pip install quantum-pecos

This will install both quantum-pecos and its dependency pecos-rslib.

For optional dependencies:

pip install quantum-pecos[all]

NOTE: The quantum-pecos package is imported like: import pecos and not import quantum_pecos.

NOTE: To install pre-releases (the latest development code) from pypi you may have to specify the version you are interested like so (e.g., for version 0.6.0.dev5):

pip install quantum-pecos==0.6.0.dev5

NOTE: Certain simulators have special requirements and are not installed by the command above. Installation instructions for these are provided here.

Rust Crates

To use PECOS in your Rust project, add the following to your Cargo.toml:

[dependencies]
pecos = "0.x.x"  # Replace with the latest version

Optional Dependencies

• LLVM version 14: Required for LLVM IR execution support (optional)

  PECOS provides an automated installer or you can install manually:

  # Quick setup with automated installer (recommended):
  cargo run -p pecos --features cli -- install llvm
  cargo build
  

  The installer automatically configures PECOS after installation.

  For detailed LLVM installation instructions for all platforms (macOS, Linux, Windows), see the Getting Started Guide.

  For full development environment setup, see the Development Setup Guide.

  Building without LLVM: If you don't need LLVM IR support:

  cargo build --no-default-features
  

Julia Package (Experimental)

PECOS also provides experimental Julia bindings. To use the Julia package from the development branch:

using Pkg
Pkg.add(url="https://github.com/PECOS-packages/PECOS#dev", subdir="julia/PECOS.jl")

Then you can use it:

using PECOS
println(pecos_version())  # Prints PECOS version

Note: The Julia package requires the Rust FFI library to be built. Currently, you need to build it locally:

1. Clone the repository
2. Build the FFI library: cd julia/pecos-julia-ffi && cargo build --release
3. Add the package locally: Pkg.develop(path="julia/PECOS.jl")

Development Setup

If you are interested in editing or developing the code in this project, see this development documentation to get started.

Simulators with special requirements

Certain simulators from pecos.simulators require external packages that are not installed by pip install .[all].

GPU-Accelerated Simulators (CuStateVec and MPS)

• CuStateVec and MPS require:
  • Linux machine with NVIDIA GPU (Compute Capability 7.0+)
  • CUDA Toolkit 13 or 12 (system-level installation)
  • Python packages: cupy-cuda13x, cuquantum-python-cu13, pytket-cutensornet

Installation: See the comprehensive CUDA Setup Guide for detailed step-by-step instructions.

Quick install (after installing CUDA Toolkit):

uv pip install quantum-pecos[cuda]

# For development with CUDA support:
make build-cuda  # Build with CUDA
make devc        # Full dev cycle (clean + build-cuda + test)
make devcl       # Dev cycle + linting

Note: When using uv or pip, install CUDA Toolkit via system package manager (e.g., sudo apt install cuda-toolkit-13), then install Python packages. Conda environments may conflict with uv/venv workflows.

Uninstall

To uninstall:

pip uninstall quantum-pecos

Citing

For publications utilizing PECOS, kindly cite PECOS such as:

@misc{pecos,
 author={Ciar\'{a}n Ryan-Anderson},
 title={PECOS: Performance Estimator of Codes On Surfaces},
 publisher = {GitHub},
 journal = {GitHub repository},
 howpublished={\url{https://github.com/PECOS-packages/PECOS}},
 URL = {https://github.com/PECOS-packages/PECOS},
 year={2018}
}

And/or the PhD thesis PECOS was first described in:

@phdthesis{crathesis,
 author={Ciar\'{a}n Ryan-Anderson},
 school = {University of New Mexico},
 title={Quantum Algorithms, Architecture, and Error Correction},
 journal={arXiv:1812.04735},
 URL = {https://digitalrepository.unm.edu/phyc_etds/203},
 year={2018}
}

You can also use the Zenodo DOI, which would result in a bibtex like:

@software{pecos_[year],
  author       = {Ciar\'{a}n Ryan-Anderson},
  title        = {PECOS-packages/PECOS: [version]]},
  month        = [month],
  year         = [year],
  publisher    = {Zenodo},
  version      = {[version]]},
  doi          = {10.5281/zenodo.13700104},
  url          = {https://doi.org/10.5281/zenodo.13700104}
}

License

This project is licensed under the Apache-2.0 License - see the LICENSE and NOTICE files for details.

Supported by