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:

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