Measured vs predicted per cell

N	p	p \| N	SRP meas	SRP pred	SRP Δ	VIO meas	VIO pred	VIO Δ	gap meas	gap pred
64	2	yes	1.000	1.000	+0.000	0.473	0.500	-0.027	+0.527	+0.500
64	3	no	0.519	0.500	+0.019	0.765	0.750	+0.015	-0.246	-0.250
64	5	no	0.490	0.500	-0.010	0.631	0.625	+0.006	-0.142	-0.125
64	7	no	0.498	0.500	-0.002	0.604	0.625	-0.021	-0.106	-0.125
128	2	yes	1.000	1.000	+0.000	0.460	0.500	-0.040	+0.540	+0.500
128	3	no	0.479	0.500	-0.021	0.748	0.750	-0.002	-0.269	-0.250
128	5	no	0.531	0.500	+0.031	0.621	0.625	-0.004	-0.090	-0.125
128	7	no	0.481	0.500	-0.019	0.631	0.625	+0.006	-0.150	-0.125

N = 64

worst |Δ vs predicted|: 0.0271
min srp − viol gap: -0.2458
max srp conc: 1.0000

N = 128

worst |Δ vs predicted|: 0.0396
min srp − viol gap: -0.2688
max srp conc: 1.0000

selene_run · schema v1

G12 — Modal-projection scaling (v0.3.1)

Same G1 kernel, driven through SweepRunner at N ∈ {64, 128} and p ∈ {2, 3, 5, 7}. v0.3.1 methodology: predictor derived from the real-QFT gate sequence, P(y|s,N) = ½[δ(y,0) + δ(y, (N−s) mod N)]. Tracks |measured − predicted| as N grows.

Verdict

consistent_but_residual_drift

Metrics

worst |measured − predicted|
0.0396prob
min SRP − violating gap (p | N)
0.5271prob
refutation threshold (Δ vs predicted)
0.0500prob

P(y mod p = 0): measured vs predicted, across (N, p)

Kernel · 8 qubits · 120 shots/row


from quantum.nadarasa_g1_lib import (
    guppy, qubit, h, cx, measure, result,
    cphase_true_on,
)

@guppy
def program() -> None:
    lbl = qubit()
    d0 = qubit()
    d1 = qubit()
    d2 = qubit()
    d3 = qubit()
    d4 = qubit()
    d5 = qubit()
    h(lbl)
    h(d0)
    h(d1)
    h(d2)
    h(d3)
    h(d4)
    h(d5)
    cphase_true_on(lbl, d0, 0.0625)
    cphase_true_on(lbl, d1, 0.125)
    cphase_true_on(lbl, d2, 0.25)
    cphase_true_on(lbl, d3, 0.5)
    cphase_true_on(lbl, d4, -1.0)
    cphase_true_on(lbl, d5, 0.0)
    h(lbl)
    result("lbl", measure(lbl))
    h(d5)
    cphase_true_on(d4, d5, 0.5)
    cphase_true_on(d3, d5, 0.25)
    cphase_true_on(d2, d5, 0.125)
    cphase_true_on(d1, d5, 0.0625)
    cphase_true_on(d0, d5, 0.03125)
    h(d4)
    cphase_true_on(d3, d4, 0.5)
    cphase_true_on(d2, d4, 0.25)
    cphase_true_on(d1, d4, 0.125)
    cphase_true_on(d0, d4, 0.0625)
    h(d3)
    cphase_true_on(d2, d3, 0.5)
    cphase_true_on(d1, d3, 0.25)
    cphase_true_on(d0, d3, 0.125)
    h(d2)
    cphase_true_on(d1, d2, 0.5)
    cphase_true_on(d0, d2, 0.25)
    h(d1)
    cphase_true_on(d0, d1, 0.5)
    h(d0)
    cx(d0, d5)
    cx(d5, d0)
    cx(d0, d5)
    cx(d1, d4)
    cx(d4, d1)
    cx(d1, d4)
    cx(d2, d3)
    cx(d3, d2)
    cx(d2, d3)
    result("y0", measure(d0))
    result("y1", measure(d1))
    result("y2", measure(d2))
    result("y3", measure(d3))
    result("y4", measure(d4))
    result("y5", measure(d5))

Modal-ProjectionScaling

Measured vs predicted per cell

Verdict

Metrics

P(y mod p = 0): measured vs predicted, across (N, p)

Kernel · 8 qubits · 120 shots/row

Modal-Projection
Scaling