# Benchmark record (rechecked 2026-08-24)

Command:

```text
moon run cmd/can-bench
```

Environment:

- Windows workspace, MoonBit `0.1.20260819`, Moonc `0.10.9+6e6c44045`.
- Default `wasm-gc` execution target.
- 2,000 iterations per workload.
- Frame workload: one extended CAN-FD frame with a 16-byte payload.
- Log workload: eight repeated event records per JSON round trip.

Observed output from the checked-out repository:

```text
moonbit-canbus benchmark; iterations=2000
frame_round_trip iterations=2000 elapsed_us=18301.7 checksum=394000
signal_pack_unpack iterations=2000 elapsed_us=2108.3999999999996 checksum=2468000
json_round_trip iterations=2000 elapsed_us=107021.5 checksum=16000
virtual_bus_steps iterations=2000 elapsed_us=873.4 tick=2000
```

These are one real local run, useful for regression comparison rather than a
cross-machine claim. Re-run the command after changing the implementation;
the benchmark intentionally prints workload, iteration count, elapsed time,
and a checksum so an optimized-away or incomplete loop is easier to detect.