This section provides performance benchmarks for all Neumann crates, measured using Criterion.rs.
# Run all benchmarks
cargo bench
# Run benchmarks for a specific crate
cargo bench --package tensor_store
cargo bench --package relational_engine
cargo bench --package graph_engine
cargo bench --package vector_engine
cargo bench --package neumann_parser
cargo bench --package query_router
cargo bench --package neumann_shell
cargo bench --package tensor_compress
cargo bench --package tensor_vault
cargo bench --package tensor_cache
cargo bench --package tensor_chain
cargo bench --package tensor_checkpoint
cargo bench --package tensor_unified
cargo bench --package neumann_serverBenchmark reports are generated in target/criterion/ with HTML visualizations.
| Component | Key Metric | Performance |
|---|---|---|
| tensor_store | Concurrent writes | 7.5M/sec @ 1M entities |
| relational_engine | Indexed lookup | 2.9us (1,604x vs scan) |
| graph_engine | BFS traversal | 3us/node |
| vector_engine | HNSW search | 150us @ 10K vectors |
| tensor_compress | TT decompose | 10-20x compression |
| tensor_vault | AES-256-GCM | 24us get, 29us set |
| tensor_cache | Exact lookup | 208ns hit |
| tensor_chain | Conflict detection | 52M pairs/sec @ 99% sparse |
| neumann_parser | Query parsing | 1.9M queries/sec |
| query_router | Mixed workload | 455 queries/sec |
| Operation | Key Metric | Performance |
|---|---|---|
| WAL writes | Durable PUT (128d embeddings) | 1.4M ops/sec |
| WAL recovery | Replay 10K records | ~400us (25M records/sec) |
All engines (RelationalEngine, GraphEngine, VectorEngine) support
optional durability via open_durable() with full crash consistency.
Benchmarks run on:
- Apple M-series (ARM64) or Intel x86_64
- Results may vary based on CPU cache sizes, memory bandwidth, and core count
For consistent benchmarking:
# Disable CPU frequency scaling (Linux)
sudo cpupower frequency-set --governor performance
# Run with minimal background activity
cargo bench -- --noplot # Skip HTML report generation for faster runs- tensor_store - DashMap concurrent storage, Bloom filters, snapshots
- tensor_compress - Tensor Train, delta encoding, RLE
- relational_engine - SQL operations, indexes, JOINs, aggregates
- graph_engine - Node/edge operations, traversals, path finding
- vector_engine - Embeddings, SIMD similarity, HNSW index
- tensor_vault - Encrypted storage, access control
- tensor_cache - LLM response caching, semantic search
- tensor_blob - Blob storage operations
- tensor_checkpoint - Snapshot create, rollback, retention
- tensor_unified - Cross-engine unified entity operations
- tensor_chain - Consensus, 2PC, gossip, sparse vectors
- neumann_parser - Tokenization, parsing, expressions
- query_router - Cross-engine query routing
- neumann_shell - Interactive CLI, WAL, snapshots
- neumann_server - gRPC server, request routing