Quiz: AI Data Center Commissioning & Burn-In (Part 13)

10 questions from Part 13 of the guide. Score instantly; every explanation links to the chapter that derives the answer.

The commissioning ladder the guide is built around is…
bronze to platinumalpha, beta, GAL1–L5: factory tests → site arrival → pre-functional → functional → integrated systems testing of the whole facilitythere is no standard ladder
Every commissioning script should carry…
a hard, pre-agreed numerical acceptance gate with a named witness and redline-on-failurea vendor signaturephotographic evidence onlyflexible targets to be fair
The guide's critique of static load-bank power acceptance is that…
it takes too longsteady-state profiles never exercise the synchronized MW-scale transients an AI cluster actually producesload banks are too expensiveload banks damage the UPS
Before proving utility-loss behavior under load, what governs the commissioning method?
a Tier label by itselfthe most dramatic available faultan approved scenario and test plan that selects the least hazardous valid evidence, with authority, prerequisites, hold points, abort criteria, communications, and restorationan unannounced utility-breaker operation
Fabric acceptance at the physical layer sets thresholds on…
pre-FEC/post-FEC bit-error rate, a link-flap budget, and a soak periodport namesswitch firmware versions onlycable colors
The cluster-level fabric gate before 'ready for production' is…
a busbw floor per collective and message size — the fabric must deliver its effective bandwidth under real collective patternscable certification alonevendor certificationone clean iperf run
How should a GPU burn-in soak duration be selected?
Use the same 72-hour minimum for every fleetUse the project/OEM/contract test plan and measured defect-discovery stopping rule; 72–168 hours is a dated practitioner exampleAlways run exactly 30 daysStop as soon as every node boots
How should SemiAnalysis's October 2024 recommendation of at least 3–4 weeks of factory high-temperature burn-in be used?
As a universal minimum burn-in durationAs an OEM RMA thresholdAs an observed universal early-failure settling periodAs a dated practitioner recommendation to evaluate against the signed OEM/project/contract plan and measured stopping rule
The energization sequence decision is about…
which switch to flip first, ceremoniallyutility paperwork orderit's fully automated everywherehow many blocks come up at once and whether each step preserves a safe retreat path
The mechanical-Cx-to-burn-in boundary should be…
decided by the GCa clean handoff at full flow on load banks — cooling proven before GPUs become the test loadwherever schedules collideskipped to save time

Check your understanding: Commissioning & Go-Live