Consistency
Cross-layer: budget and key rate must model the untrusted detector once
8/8 passed in 1ms
| Test | What it does | Result |
|---|---|---|
test_budget_silent | budget.assemble takes no vel or trusted argument, budget.detector does not exist, and an assembled budget has no detector row. | ✅ pass |
test_double_count | Adding mu*v_el/T to a hardware budget by hand lowers the Link rate by over 1e-3 on both detectors. | ✅ pass |
test_eta_referral | The untrusted homodyne rate matches xi + v_el/(eta*T) to 1e-12 and not xi + v_el/T. | ✅ pass |
test_iab_invariant | Trusted and untrusted I_AB agree to 1e-12 over the grid on both detectors. | ✅ pass |
test_merged_channel | Untrusted mode equals a trusted run on the merged channel T -> etaT, xi -> xi + muv_el/(eta*T), mu = 1 homodyne and 2 heterodyne, over the (T, eta, v_el, V_A) grid: Laudenbach & Pacher arXiv:1904.01970 Sec. 2. | ✅ pass |
test_mu_doubling | The untrusted heterodyne rate matches the mu = 2 fold to 1e-12 and not mu = 1 (Laudenbach arXiv:1703.09278 Eqs. (9.93), (9.94)). | ✅ pass |
test_single_count | A hardware budget through an untrusted homodyne Link matches the merged channel to 1e-12. | ✅ pass |
test_untrusted_costs | The untrusted rate never exceeds the trusted one, equal at eta = 1, v_el = 0. | ✅ pass |