baseline-native

Native In-Process Benchmark

This run uses the Rust ative_inprocess_benchmark bench and loads each engine DLL directly in the measured process. It does not use the historical managed .NET/PInvoke benchmark host.

• Track A: luna_pinyin, Yune versus librime 1.17.0.
• Track B: jyut6ping3_mobile, Yune product path before/after.
• Track A inputs: $TrackAInputs.