More of the same hardware spent time generating
Distributed AI is often framed as a search for bigger clusters. This benchmark started from the opposite constraint: two compact laptops, an open model workload, and no interest in changing what the model is meant to produce.
On our internal test, the unoptimized path landed near 50 tokens per second. With TDCC’s coordination layer in place, the same Qwen 3.8 27B workload reached 80 tokens per second. That is a 60% improvement in observed generation throughput.
~50
tok/s on the baseline path
80
tok/s on the TDCC-coordinated path
+60%
observed generation throughput
We did not alter model weights, swap in a smaller model, or aim for a different output behavior. The work was in the path around generation:
- coordinating when work moves,
- reducing avoidable idle time,
- and keeping the stream moving across both devices.
50 → 80 tok/s, with the model’s intended behavior unchanged
Internal benchmark result
Same two-device setup. Same Qwen 3.8 27B workload. The comparison is about keeping the generation path productive, not modifying the model itself.
Three ways to leave less useful time on the table
Diagram 03 · the high-level work
Coordinate the work.
Give available devices a shared, orderly role in the generation path instead of treating them as isolated endpoints.
Reduce avoidable waiting.
Keep each handoff purposeful so capable hardware spends less time waiting for the next useful unit of work.
Keep generation streaming.
Once output starts, protect continuity. The aim is one coherent stream for the user, not a visible sequence of device boundaries.
What we are testing
Engineering principle
Useful compute is not only about having more hardware. It is about helping the hardware you already have spend more time doing useful work.