Qwen3-14B 4095+1-token inference smoke — allocator reserved

Qwen3-14B 4095+1-token inference smoke — allocator reserved: 7.89 GiB on a 12 GB card.

Single reading

8,472,494,080 bytes

MEASURED recorded 21 September 2026

Protocol

Test definition

GPU memory telemetry from the project's watcher / allocator counters. After memory-efficient attention repair; dedicated 8,669,196,288.

Limits

Limitations

  • Sampled gauges; sub-sample transients not exhaustively proven. Physical card is 12,227 MiB; allocator and dedicated readings are different counters.

Log

Recorded values

Date Value Evidence Samples Source
21 September 2026 8,472,494,080 bytes MEASURED docs/ledger/FINAL_LM_BOUNDED_FEASIBILITY_DECISION.md

Source paths refer to the private JARVIS repository and runtime records. They are listed for audit and are not published.