OUTDATED / HISTORICAL ARCHIVE — This report predates the fresh three-repeat campaign. Its timings, MFU inputs and optimization conclusions are superseded. Open the current evidence report ↗

WORLD MODEL PERFORMANCE DOSSIER · H200-NORMALIZED THEORY

LingBot-World 2.0

01 · THEORY

Compute, communication, and H200 sizing

interactive model · target > 12 fps

The explorer starts at 13 fps—the first whole-frame target above 12. Switch the shared matched resolution, target, real-run efficiency calibration, parallel strategy, and GPU count. The model recomputes token geometry, FLOPs, collective traffic, block latency, and the minimum power-of-two H200 count. Values in this section are projections; measured values appear in the next section.

Selected configuration: compute vs communication

Per-forward FLOP composition

Scaling and communication by H200 count

02 · EVIDENCE

Real runs compared with theory

matched H×W · real U1 · p50 / p95

480×832 vs 704×1280 at one timing boundary

03 · RUNTIME

Where real run time goes

settings · profiler · decode

Matched-resolution block boundary

Observed time by component and setting

04 · ACTION

Bottlenecks and required optimization

diagnosis → next move