ATLAS-Q v0.5.0

ATLAS-Q v0.5.0

GPU-accelerated quantum tensor network simulator with adaptive MPS, custom Triton kernels, and verified benchmarks.

Installation

pip install atlas-quantum[gpu]

Or try the https://colab.research.google.com/github/followthesapper/ATLAS-Q/blob/ATLAS-Q/ATLAS_Q_Demo.ipynb in Google Colab.

Key Features

- Adaptive MPS: 626,000× memory compression vs full statevector (30 qubits)
- Custom Triton Kernels: 1.5-3× speedup on gate operations
- Stabilizer Backend: 20× speedup for Clifford circuits
- Period-Finding: Shor's algorithm with O(1) memory for periodic states
- NISQ Simulation: Noise models with Kraus operators
- Variational Algorithms: VQE, QAOA with GPU optimization
- Time Evolution: TDVP for Hamiltonian dynamics
- 2D Circuits: Automatic SWAP insertion for grid topologies

Performance

- 77,304 ops/sec gate throughput (GPU)
- 30 qubits in 0.03 MB (vs 16 GB statevector)
- All 7/7 benchmark suites passing

What's Included

- Python package for PyPI (atlas-quantum)
- Docker images (GPU + CPU)
- Interactive Jupyter notebook demo
- Complete documentation with verified examples
- GitHub Actions CI/CD workflows

Documentation

- https://github.com/followthesapper/ATLAS-Q/blob/ATLAS-Q/docs/COMPLETE_GUIDE.md
- https://github.com/followthesapper/ATLAS-Q/blob/ATLAS-Q/docs/FEATURE_STATUS.md
- https://github.com/followthesapper/ATLAS-Q/blob/ATLAS-Q/ATLAS_Q_Demo.ipynb
- https://followthesapper.github.io/ATLAS-Q/

Quick Example

from atlas_q import get_quantum_sim

QCH, _, _, _ = get_quantum_sim()
sim = QCH()
factors = sim.factor_number(221)
print(f"221 = {factors[0]} × {factors[1]}")  # 221 = 13 × 17

Requirements

- Python 3.9+
- PyTorch 2.0+
- CUDA 11.8+ (for GPU acceleration)
- Triton 2.0+ (optional, for custom kernels)

License

MIT License