======================================================================== LOCK-FREE FLIP — FROZEN ANALYSIS (ledger §8) scipy available: True ======================================================================== === Throughput inference [variant A — disjoint (PRIMARY)] (write_ops_per_sec) === CONTROL : n=30 mean=1,986.9 sd=201.6 TREATMENT: n=30 mean=14,969.7 sd=3,109.7 diff (T-C) = 12,982.8 ops/s (+653.42%) 95% Welch CI [11,819.6, 14,146.0] Welch t = 22.8195 df = 29.24 p(two-sided) = 3.4958e-20 Cohen's d = 5.8920 (pooled) Mann-Whitney U = 900.0 p(two-sided) = 3.0199e-11 === S1 — checkpoint pause (veto: TREATMENT ≤ CONTROL) === [S1 PRIMARY] writer-observed p999 (us): control=258737.6 treatment=10269.4 [secondary 'ALSO'] checkpointer-side mean (us): control=164244.9 treatment=193765.3 [secondary 'ALSO'] checkpointer-side p99 (us): control=272617.4 treatment=369489.7 === S2 — write tail p50/p99/p999 (veto: T ≤ 1.10×C on p99/p999) === write_p50_us: control=173.8 treatment=417.3 ratio(T/C)=2.401 write_p99_us: control=145326.0 treatment=3676.4 ratio(T/C)=0.025 write_p999_us: control=258737.6 treatment=10269.4 ratio(T/C)=0.040 === S3 — peak RSS (single-arm-per-process pass; veto: T ≤ 1.25×C) === peak RSS (KiB): control=3,471,144 treatment=1,018,664 ratio(T/C)=0.293 === Throughput inference [variant B — contended (S4)] (write_ops_per_sec) === CONTROL : n=30 mean=2,439.1 sd=124.1 TREATMENT: n=30 mean=14,806.0 sd=741.4 diff (T-C) = 12,366.9 ops/s (+507.02%) 95% Welch CI [12,086.8, 12,647.0] Welch t = 90.1040 df = 30.62 p(two-sided) = 1.055e-38 Cohen's d = 23.2647 (pooled) Mann-Whitney U = 900.0 p(two-sided) = 3.0199e-11 S4 cas_retries (treatment, per round): mean=127.4 max=228 === §8 DECISION RULE === -> BRANCH 1: PROCEED. throughput +12,982.8 ops/s (95% CI [11,819.6,14,146.0]), p=3.496e-20, d=5.89≥0.8, no S1/S2/S3/S4 regression ======================================================================== VERDICT: branch 1 = PROCEED EVIDENCE: throughput +12,982.8 ops/s (95% CI [11,819.6,14,146.0]), p=3.496e-20, d=5.89≥0.8, no S1/S2/S3/S4 regression ========================================================================