Expand description
Scanline execution engine.
Owns the query execution loop: it takes an aligned buffer from
scanline-storage, splits it into per-core chunks, and drives the predicate
kernels in scanline-simd across pinned worker threads.
Phase 0 scaffold: the chunk scheduler and core-affinity pinning land in Phase 1. What exists here now is the single-threaded execution path, which is already enough to saturate a meaningful fraction of memory bandwidth and is the baseline every parallel result must beat.
Structs§
- Scan
Result - Outcome of one scan, carrying enough context to be reported honestly.
Functions§
- scan_
greater_ than - Runs a greater-than predicate over an aligned
i64buffer.