2025 practitioner example for GPU node burn-in/soak; not a universal acceptance minimum or maximum
72–168 hrrecommendation
| Value kind | recommendation — Recommendations are design guidance, not measured facts; validate them against the project boundary conditions. |
|---|---|
| Scope | Together AI and ClusterMAX described a 72–168 hr range in 2025 practitioner guidance. The project/OEM/contract test plan sets duration, intensity, statistical stopping rule, pass/re-soak gate, and vendor disposition. |
| As of | 2025 |
| Source | Together AI seven-phase guide; ClusterMAX 2.0 |
| Review | checking…review by 2026-11-23 · standard cadence |
| Recorded changes | last 2026-07-10 · 2 revisions tracked |
| Claim id | gpu-node-burn-in-soak-window-3-day-minimum-to-7 |
Where the guide uses it
- 10.5 Provisioning, Bring-Up & Infrastructure as Code
- 13.2 Documentation, Scripts & Acceptance Test Plans
- 13.8 GPU Node Burn-In, Diagnostics & Stress Validation
- 14.3 Component Failure Modes, Failure Rates & Fleet Reliability Data
← Full numbers register — every date-stamped figure in the guide, with revision history.