The Graph Scaling Challenge
Reference: This page is a source-based technical synthesis of the LaTeX chapter cited below. For exact syntax, availability or ABI signatures, verify the versioned source, manifest and executable tests.
Scope and source boundary
Scaling requires workload-specific indexes, candidate selection and bounded evaluation while preserving exactly the same semantic authority.
Indexing accelerates access to semantic structures but does not replace semantic authority. The neuro-symbolic architecture distinguishes source, assembled model, compiled semantics, execution state and derived numerical indexes. Each index can be rebuilt or invalidated without silently changing the meaning of the underlying graph.
Engineering rules
- Describe which authority owns the indexed fact.
- Track invalidation when declarations, relations, or embeddings change.
- Treat approximate ranking as candidate retrieval, not logical proof.
- Benchmark the relevant workload without inventing universal speedup claims.
Chapter outline (original LaTeX headings)
- Effect We Want
- Canonicalization Before Expansion
- Primary Selective Indexes
- Join Planning
- Delta and Semi-Naive Evaluation
- Branch State Management
LaTeX provenance
Primary chapter: logicells-neuro-symbolic-architecture-guide/chapters/performance-execution-architecture.tex.