A paid production workload on the reference fleet, measured on GeForce RTX 4080 and GeForce RTX 5090 on the same days. One workload, its own cells — never averaged with any other workload, because cross-workload aggregation is the substitution error this project exists to expose.
| DAY | GOODPUT GeForce RTX 4080 / GeForce RTX 5090 | ATTEMPTS | COMPLETIONS | CARDS |
|---|---|---|---|---|
| 2026-08-21 | 0.39× | 42 / 155 | 31 / 149 | 2 / 6 |
| 2026-08-22 | 0.25× | 25 / 187 | 12 / 178 | 2 / 6 |
| 2026-08-23 | 0.48× | 55 / 157 | 53 / 151 | 2 / 6 |
| 2026-08-24 | 0.48× | 40 / 91 | 37 / 78 | 2 / 4 |
| 2026-08-26 | 0.35× | 79 / 195 | 69 / 188 | 2 / 6 |
| 2026-08-27 | 0.36× | 131 / 178 | 120 / 171 | 2 / 6 |
| 2026-08-29 | 0.36× | 54 / 31 | 49 / 30 | 2 / 3 |
| 2026-09-01 | 0.29× | 79 / 68 | 71 / 65 | 2 / 2 |
| 2026-09-03 | 0.51× | 66 / 27 | 59 / 27 | 2 / 4 |
The measured pair page for these deployments: side-by-side cells. Downloads: ratio cells, value board, the rental prices these were costed at.