data/README.md

# Data Directory — Pre-Computed Results

This directory contains the five output `.pkl` files generated by running the QR-SPPS
five-notebook pipeline on the **Fujitsu QSim A64FX cluster** (Fujitsu QARP v0.4.4).

## Files

| File | Size (approx.) | Contents |
|---|---|---|
| `QRSPPS_hamiltonians.pkl` | ~10 KB | 40-qubit Hamiltonian, exact sub-network eigenvalues (12q, 16q), spectral gap |
| `QRSPPS_vqe_results.pkl` | ~32 MB | VQE ground state energies, stress distributions for all 40 nodes, quantum advantage map, MC baseline |
| `QRSPPS_policy_results.pkl` | ~46 KB | ADAPT-VQE gradients for 6 policies, energy changes, node-level stress delta matrix |
| `QRSPPS_dosqpe_results.pkl` | ~13 KB | Survival amplitude trajectory, density of states, Boltzmann tail risk curves, cascade dynamics |
| `QRSPPS_scaling_results.pkl` | ~5 KB | 12–30q scaling benchmarks, exponential fit parameters, pipeline summary |

## Verification

Every numerical result in the paper is verifiable with a single Python command:

```python
import pickle

# Example: verify VQE zero error
vqe = pickle.load(open("QRSPPS_vqe_results.pkl", "rb"))
assert abs(vqe["vqe_energy_30q"] - (-33.5198)) < 1e-4
assert abs(vqe["vqe_energy_40q"] - (-44.6931)) < 1e-4
assert vqe["vqe_error"] == 0.0
print("✅ VQE results verified: zero error confirmed.")

# Example: verify policy results
pol = pickle.load(open("QRSPPS_policy_results.pkl", "rb"))
assert abs(pol["stockpile_delta_e40"] - (-7.4505)) < 1e-4
print("✅ Policy results verified: Stockpile ΔE[40q] = -7.4505")
```

## Provenance

These files were generated on:
- **Hardware:** Fujitsu QSim FX700, ARM A64FX compute nodes
- **QARP version:** Fujitsu QARP v0.4.4 (Production Build)
- **Qulacs:** 0.6.12 (A64FX-optimised MPI kernel, SVE-accelerated)
- **Python:** 3.12 via pyenv
- **MPI allocation:** 4 nodes × 12 tasks/node = 48 MPI ranks
- **Account:** Group A (g140-user1)
- **Challenge:** Fujitsu Quantum Simulator Challenge 2025-26