Independent P8 versus P1 polling replication
The P8 host-polling gain replicates on eight unseen headers without exceeding the registered 0.20-second polling interval.
Diagnostic work, exact algebra and local capabilities are not treated as end-to-end mining advantage.
What was tested?
The P8 host-polling gain replicates on eight unseen headers without exceeding the registered 0.20-second polling interval.
Why the test is meaningful
A small orchestration effect must replicate with frozen binaries and must satisfy both throughput and operational-latency gates.
Rwall=wall throughput(P8)/wall throughput(P1)gate R≥1.002 and CI_low>1poll interval(P8)<0.20 sHow it was tested
Freeze P1/P8, generate eight unseen headers and run 64 balanced ABBA/BAAB measurements.
What happened
P8 won 8/8 headers with ratios 1.002628–1.003139. Wall ratio was 1.002858 and 95% interval 1.002725–1.002991. P8 polling interval was 0.162203 seconds versus 0.020333 for P1.
Exactness and statistical controls
Exactness and compiled-depth checks passed; no reselection or threshold adaptation occurred.
What the result means
P8 is a confirmed local throughput capability within the latency gate, pending energy/thermal confirmation.
Limitations
- Detection latency is eight times longer than P1.
- Energy remains a separate gate.
- The result is GPU-host orchestration, not SHA probability advantage.
Evidence trail
Use frozen binaries, eight unseen headers, equal work and measure both wall throughput and interval.
Canonical variants
CANONICAL-EXP-150PREREGISTERED-CAMPAIGNSEALED-AUDITSource: internally audited canonical reports. Local filesystem structure, private headers and operational identifiers are excluded from publication.