eng-0086
window recorded
citable URL: https://halobench.com/records/eng-0086/ — this address never moves; the anchor /records/#eng-0086 keeps resolving
27.57 Whmean 183.49 W − 10.1 W idle floor (box-idle-all-empty) = 173.39 W attributable
window 2026-08-15T11:25:19Z → 2026-08-15T11:34:20Z · provenance recorded
method wall-meter · aihydra energy sensor via Home Assistant (counter-difference on sensor.hardware_ai_hydra_energy) · no contention
covers run-0276
EOS-cliff x q8_0-KV probe, n_max=7 (run-0276, cfg-0074) — a cliff-detector guard run (15 generation requests, 0 degenerate), not a reward-scored capability run; wh_per_task/pence_per_correct_answer left null for the same reason as eng-0085. Wh/request = 27.5745/15 = 1.84 — higher than n_max=3's 1.28 Wh/request (eng-0085), consistent with a larger draft window costing more compute per accepted step. Counter samples: start 19.2474132031202276 kWh (2026-08-15T11:25:19Z, exact — shared boundary with eng-0085's end, the two runs are back-to-back on one server session), end 19.2749877423047976 kWh (2026-08-15T11:34:19Z, 1s before window close — nearest recorded tick). Retroactive join computed 2026-08-15, same day as the run.
Wh, never joules — read as a cumulative-counter difference across the window; delta over the idle baseline is the only figure that means anything