Skip to content

ATLAS-Q v0.6.0

Choose a tag to compare

View all tags
@followthesapper followthesapper released this 27 Oct 20:17
· 33 commits to ATLAS-Q since this release
v0.6.0
0b9ce0b

ATLAS-Q v0.6.0 - Molecular Chemistry & Advanced Tensor Networks

Release Date: October 2025

We're excited to announce ATLAS-Q v0.6.0, featuring quantum chemistry integration, graph optimization, and advanced tensor network capabilities. This release adds
38 new tests, bringing total integration test coverage to 46/46 passing ✅.

🎯 What's New

Priority 1: Quantum Chemistry & Graph Optimization

Molecular Hamiltonians - Full quantum chemistry support

  • PySCF integration for electronic structure calculations
  • Jordan-Wigner transformation for fermion-to-qubit mapping
  • Support for H₂, LiH, H₂O, and custom molecular geometries
  • Seamless VQE integration for ground state energy calculations
  • 4/4 tests passing

MaxCut QAOA - Graph optimization Hamiltonians

  • Automatic MaxCut Hamiltonian builder for QAOA
  • Weighted and unweighted graph support
  • Automatic edge normalization for undirected graphs
  • Perfect for combinatorial optimization problems
  • 4/4 tests passing

Priority 2: Advanced Tensor Network Features

Circuit Cutting - Partition large circuits

  • Min-cut graph partitioning algorithm
  • Entanglement analysis and visualization
  • Automatic sampling overhead calculation
  • Run large circuits on small quantum computers
  • 7/7 tests passing

PEPS 2D Tensor Networks - True 2D quantum simulation

  • Projected Entangled Pair States for 2D grid topologies
  • Boundary MPS contraction for measurements
  • Optimized for shallow circuits (Google Sycamore topology)
  • 10/10 tests passing

Distributed MPS - Multi-GPU scaling

  • Bond-parallel domain decomposition
  • Single-node multi-GPU support (data & model parallelism)
  • Ready for multi-node scaling (future)
  • 10/10 tests passing

cuQuantum Backend - NVIDIA acceleration (optional)

  • Integration with cuQuantum 25.09.1
  • 2-10× speedup on large tensor operations
  • Transparent fallback to PyTorch
  • Accelerated SVD, QR, and tensor contractions
  • 11/11 tests passing

🚀 New Features

Command-Line Interface

New CLI with intuitive commands

atlas-q --help # Show help
atlas-q factor 221 # Factor integers
atlas-q benchmark # Run all 46 tests
atlas-q info # System information
atlas-q demo # Interactive demo

Updated Demo Notebook

ATLAS_Q_Demo.ipynb now includes:

  • Molecular chemistry example (H₂ with VQE)
  • MaxCut graph optimization
  • Circuit cutting demonstration
  • PEPS 2D network creation
  • cuQuantum acceleration example

All examples tested and working on CPU & GPU ✅

📊 Performance & Testing

  • 46/46 integration tests passing (up from 8)
  • 77K+ ops/sec gate throughput (unchanged)
  • 626,000× memory compression (unchanged)
  • 20× speedup on Clifford circuits (unchanged)
  • CI enabled for Dev branch testing

📚 Documentation

Completely Updated:

  • ✅ README.md - Added v0.6.0 features, CLI section
  • ✅ COMPLETE_GUIDE.md - Added Section 3.9 (Priority 2 features) with API references
  • ✅ WHITEPAPER.md - Added Section 4.7 for advanced tensor networks
  • ✅ OVERVIEW.md - Updated capabilities and version
  • ✅ RESEARCH_PAPER.md - Added v0.6.0 to appendix
  • ✅ FEATURE_STATUS.md - Moved Priority 2 from "⚠️ Partially Implemented" to "✅ Fully Tested"
  • ✅ ATLAS_Q_Demo.ipynb - Added 5 new working examples

🔧 Installation

From PyPI (Coming Soon)

pip install atlas-quantum # Currently at v0.5.0

v0.6.0 will be published soon

From Source (Recommended for v0.6.0)

git clone https://github.com/followthesapper/ATLAS-Q.git
cd ATLAS-Q
git checkout ATLAS-Q # or Dev branch
pip install -e .[gpu]

For molecular chemistry

pip install pyscf

For cuQuantum acceleration (optional)

pip install cuquantum-python

Docker

GPU version (recommended)

docker pull ghcr.io/followthesapper/atlas-q:cuda

CPU version

docker pull ghcr.io/followthesapper/atlas-q:cpu

💡 Quick Examples

Molecular Chemistry

from atlas_q import get_mpo_ops, get_vqe_qaoa

mpo = get_mpo_ops()
H = mpo['MPOBuilder'].molecular_hamiltonian_from_specs(
molecule='H2', basis='sto-3g', device='cuda'
)

vqe_mod = get_vqe_qaoa()
vqe = vqe_mod['VQE'](H, vqe_mod'VQEConfig')
energy, params = vqe.run()
print(f"Ground state energy: {energy.real:.6f} Ha")

Graph Optimization

from atlas_q import get_mpo_ops, get_vqe_qaoa

mpo = get_mpo_ops()
edges = [(0, 1), (1, 2), (0, 2)] # Triangle graph

H = mpo['MPOBuilder'].maxcut_hamiltonian(edges, device='cuda')

qaoa_mod = get_vqe_qaoa()
qaoa = qaoa_mod['QAOA'](H, n_layers=2)
cost, params = qaoa.run()
print(f"MaxCut value: {-cost:.4f}")

Circuit Cutting

from atlas_q import get_circuit_cutting

cutting = get_circuit_cutting()
graph = cutting'CouplingGraph'

for i in range(7):
graph.add_two_qubit_gate(i, i+1)

config = cutting'CuttingConfig'
cutter = cutting'CircuitCutter'
partitions = cutter.cut(graph, n_partitions=2)

⚠️ Breaking Changes

None. All v0.5.0 code remains fully compatible.

🐛 Known Issues

  • cuQuantum tests run in fallback mode in CI (no CUDA toolkit on GitHub runners)
  • Multi-node distributed MPS requires additional setup (torch.distributed)
  • PyPI package still at v0.5.0 (v0.6.0 package coming soon)

📈 What's Next (v0.7.0)

  • Qiskit/Cirq circuit import adapters
  • Additional tutorial notebooks
  • PyPI package update to v0.6.0
  • Multi-node distributed MPS testing
  • Expanded molecular chemistry examples

🙏 Acknowledgments

  • PyTorch team for GPU infrastructure
  • PySCF team for quantum chemistry integration
  • NVIDIA cuQuantum team for tensor acceleration
  • Community for feature requests and testing

📞 Support

Full Changelog: v0.5.0...v0.6.0

License: MIT | Python: 3.8+ | PyTorch: 2.0+

ATLAS-Q v0.6.0: Production-ready quantum tensor network simulator with molecular chemistry, graph optimization, and advanced tensor network capabilities. All
46/46 tests passing ✅

Assets 2
Loading