GPUDECK · WORKLOAD EVIDENCE

krea2_turbo_fp8_scaled

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.

14 ratio cell(s) · window 2026-08-21 → 2026-09-04 · page generated 2026-09-23 05:21 UTC

Dated ratio cells

DAYGOODPUT GeForce RTX 4080 / GeForce RTX 5090ATTEMPTSCOMPLETIONSCARDS
2026-08-210.40×455 / 937416 / 9072 / 6
2026-08-220.30×474 / 1364419 / 13222 / 6
2026-08-230.42×628 / 1131582 / 10992 / 6
2026-08-240.58×60 / 19259 / 1852 / 6
2026-08-260.42×563 / 1124522 / 10782 / 6
2026-08-270.31×434 / 1264393 / 12282 / 6
2026-08-280.66×134 / 22127 / 192 / 4
2026-08-290.36×346 / 355326 / 3432 / 6
2026-08-300.57×30 / 4026 / 382 / 3
2026-08-310.52×789 / 57728 / 542 / 5
2026-09-010.44×502 / 48457 / 472 / 3
2026-09-020.58×29 / 1629 / 152 / 5
2026-09-031.32×671 / 181655 / 1722 / 5
2026-09-040.55×33 / 4330 / 432 / 5
Confounds travel with the number. † marks a host confound. Attempts share a card, host and queue - replication is physical cards and card-days, never attempt count.

The measured pair page for these deployments: side-by-side cells. Downloads: ratio cells, value board, the rental prices these were costed at.