Skip to main content

Module closures

Module closures 

Source
Expand description

Phase 2: compiled u64-only kernels with flat buffer evaluation.

Four monomorphic kernel types, each produced by a distinct compiler path. No runtime branching for optimization strategy — the eval loop is baked in at construction time.

TypePush (per-node skip)Pull (cone guard)
CompiledKernelRaw
CompiledKernelPushyes
CompiledKernelPullyes
CompiledKernelPushPullyesyes

Structs§

CompiledKernelPull
The closure tier with the cone guard: an output whose cone no changed input reaches is not recomputed.
CompiledKernelPush
The closure tier with per-step skipping: a changed input invalidates its dependents through the plan, and a current step is skipped.
CompiledKernelPushPull
The closure tier with per-step skipping and the cone guard.
CompiledKernelRaw
The closure tier with no provenance: every evaluation runs every step.