Quantum Simulators

Use this guide to choose a simulator quickly and consistently.

Selection Rule

• Default: UnitaryLab.
• If UnitaryLab is unavailable: Qiskit.
• If the task is differentiable hybrid optimization: PennyLane.

Installation Note

• For any installation path that uses uv, install it with python -m pip install -U uv first, then verify with uv --version.
• If pip-based install is unavailable, use the platform-specific installer commands in the UnitaryLab guide.

UnitaryLab (Recommended)

Reference: ./unitarylab/SKILL.md

Best when:

• Learning and teaching quantum fundamentals.
• Building small to medium algorithm demos quickly.
• Working on PDE and Schrodingerization style workflows.

Strengths:

• Simple circuit interface.
• Lightweight local workflow.
• Good fit for educational and algorithm prototypes.
• Available for Windows, macOS, and Linux via platform-specific wheels.

Wheel Files (Python 3.11 required)

OS	Wheel filename
Windows x64	unitarylab-0.1.0-cp311-cp311-win_amd64.whl
macOS (arm64)	unitarylab-0.1.0-cp311-cp311-macosx_11_0_arm64.whl
Linux x86-64	unitarylab-0.1.0-cp311-cp311-linux_x86_64.whl

All wheels are located in ./unitarylab/dist/. Use the one matching the user's OS and architecture.

Qiskit

Reference: ./qiskit/SKILL.md

Best when:

• You need richer ecosystem features or noise models.
• You plan to move toward IBM hardware workflows.
• UnitaryLab is not installed in the current environment.

PennyLane

Reference: ./pennylane/SKILL.md

Best when:

• You need differentiable circuits.
• You are implementing VQE, QAOA, or QML pipelines.
• You need tight integration with PyTorch or TensorFlow.

Required Practice for Algorithm Examples

• Always run on a simulator, not only circuit construction.
• Always report at least one validation signal:
  • statevector or measurement probabilities,
  • comparison against a classical/theoretical expectation.
• Keep first examples minimal and executable.