mirror of https://github.com/luxfi/consensus.git synced 2026-07-27 02:57:46 +00:00

Files

T

zeekayandHanzo Dev 7197145afc docs: retract fabricated throughput numbers (honesty pass)

Remove impossible/mislabeled per-second figures from consensus docs: the
"3.9B blocks/sec" / "6.6B votes/sec" u8-overflow artifacts (real C-FFI ≈21K
votes/sec per the CI README), the 20.26M/11.5M "TPS" that are ZAP wire-message
microbenches (relabeled, not consensus finality), stub-SDK "sub-second finality"
claims, and downgraded "production-ready" labels on SDKs the CI audit documents
as stubs. Genuinely-measured go-test numbers left intact. Fixed stale
LUX_CONSENSUS_PQ_MODE → CONSENSUS_PQ_MODE. Docs only, no Go logic touched.

Co-authored-by: Hanzo Dev <dev@hanzo.ai>

2026-07-02 08:53:06 -07:00

6.0 KiB

Raw Permalink Blame History

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

[1.21.0] - 2025-11-06

Added

Multi-Language SDK Implementation

C implementation: Complete C library with 33 comprehensive tests (445 LOC test code)
Rust FFI bindings: Safe wrapper with 4 core tests + Criterion benchmarks (875 LOC)
Python SDK via Cython: Pythonic API with comprehensive test suite (674 LOC)
C++ with MLX: GPU-accelerated implementation for Apple Silicon
Go CGO integration: Seamless C/Go interop for performance-critical paths

Photon/Emitter Refactoring

Replaced Sampler/Sample pattern with light-themed Emitter/Emit architecture
Implemented luminance tracking for performance-based node selection
- Range: 10-1000 lux (twilight → bright daylight)
- Base: 100 lux (office lighting)
- Dynamic brightness adjustment: +10% on success, -10% on failure
Performance-based weighting using brightness values for node selection

Quantum-Resistant Integration

OP Stack quantum finality integration example (examples/op_stack_quantum_integration.go)
Post-quantum cryptographic proofs:
- ML-DSA-65 digital signatures
- ML-KEM-1024 key encapsulation
- Quantum-resistant Merkle tree implementation
Layer 2 rollup integration with quantum-resistant finality guarantees

Examples and Documentation

7 progressive tutorial examples:
- 01-simple-bridge: Cross-chain transfer basics
- 02-ai-payment: AI-powered payment validation
- 03-qzmq-networking: Quantum-secure ZMQ messaging
- 04-grpc-service: gRPC service integration (planned)
- 05-python-client: Python SDK usage examples
- 06-typescript-integration: TypeScript SDK (planned)
- 07-ai-consensus: Full AI consensus orchestration
Package-level documentation (doc.go) with comprehensive API docs
~8,000 lines of new documentation across examples and guides

Changed

Architecture Improvements

Moved 1,631 lines of marketplace code from ai/ to examples/ai-marketplace/
Consolidated duplicate structures (configs, interfaces, context)
Net reduction of 672 lines through deduplication
Clean, DRY, orthogonal package layout

Test Coverage Improvements

AI package: 37.1% → 74.5% coverage (+37.4pp)
Core consensus: 432/580 lines tested
Removed untestable blockchain-dependent code to examples
High-coverage packages maintained:
- utils/codec: 100%
- version: 100%
- protocol/flare: 95.7%
- protocol/focus: 91.1%
- protocol/horizon: 88.9%
- protocol/wave: 83.3%

Performance

AI Consensus Benchmarks (Apple M1 Max, Go 1.24.5):

Model decisions: 1.5 μs per decision (660K decisions/sec)
Simple model learning: 628 ns per training example
Feature extraction: 37 ns (zero allocations)
Sigmoid operations: 5.6 ns (179M ops/sec)
Memory efficiency: 912 bytes per decision, 18 allocations

C Implementation:

33/33 tests passing (100% pass rate)
1000 blocks processed in < 0.001 seconds
Zero-copy operations where possible
Minimal memory footprint

Rust Implementation:

4/4 tests passing (100% pass rate)
Zero-cost abstractions
Memory safety guarantees

Multi-Language Status:

Language	Status	Tests	Test LOC
Go	Production	Full suite	Core
C	Production	33 tests	445
Rust	Production	4 tests	875
Python	Research	Comprehensive	674
C++/MLX	Development	In progress	-

Deprecated

Sampler/Sample APIs in photon package (still functional, use Emitter/Emit for new code)

Fixed

Build compatibility across all language implementations
Test coverage reporting for AI package
Documentation gaps in protocol packages

Security

Quantum-resistant cryptography integration for Layer 2 rollups
Post-quantum signature schemes (ML-DSA)
Post-quantum key encapsulation (ML-KEM)

[1.17.0] - 2025-01-12

🎉 Major Release: Multi-Language SDK with 100% Test Parity

This release introduces complete multi-language support for the Lux consensus engine with C, Rust, Python, and Go implementations, all achieving 100% test parity and comprehensive benchmarks.

✨ Features

Multi-Language Implementation

C Implementation: High-performance native C library with optimized hash tables
Rust FFI Bindings: Safe Rust wrapper with zero-cost abstractions
Python SDK (Cython): High-level Python bindings with Pythonic API
Go CGO Integration: Conditional compilation with seamless switching

Quantum-Resistant Features

OP Stack Integration with ML-DSA-65 and ML-KEM-1024
Quantum-resistant Merkle trees
Example implementation provided

Performance Achievements

Per-second throughput figures from this release were retracted by the 2025-11-10 SDK audit: the C/Rust/Python/Go SDKs measure data-structure insertion (or the Rust u8 overflow bug), not consensus. The honest consensus-bound rate is ≈21K votes/sec (C FFI). See .github/workflows/README.md and the "Honest Assessment" block in LLM.md.

📊 Test Coverage

100% test parity across all 4 implementations
15 test categories with comprehensive coverage
All performance targets exceeded

🔧 Technical Improvements

Custom hash table for O(1) operations
Thread-safe with read-write locks
Vote caching for performance
Memory-efficient implementations

📦 Installation

See README.md for detailed installation instructions for each language.

📈 Benchmarks

Run benchmarks with:

C: make benchmark
Rust: cargo bench
Python: python3 benchmark_consensus.py
Go: go test -bench=.

6.0 KiB Raw Permalink Blame History