Files
dex/CHANGELOG.md
2025-12-11 17:50:19 -08:00

2.6 KiB

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.

[1.2.1] - 2025-12-11

Added

  • Q-Chain quantum finality verifier integrated into Oracle
  • MLX (Apple Silicon) FIX protocol benchmark support
  • oracle.SetVerifier(v) to attach Q-Chain verifier
  • oracle.VerifiedPrice(sym) returns price with finality status
  • VerifiedData type includes quantum finality proof

Fixed

  • Fix double-close panic in QChainVerifier
  • Fix orderbook insert infinite loop

Performance (FIX Protocol Benchmarks)

Engine NewOrderSingle ExecutionReport MarketDataSnapshot Avg Latency
Pure Go 163K/sec 124K/sec 332K/sec 33.5 μs
Hybrid Go/C++ 167K/sec 378K/sec 616K/sec 17.3 μs
Pure C++ 444K/sec 804K/sec 1.08M/sec 8.2 μs
Rust 484K/sec 232K/sec 586K/sec 11.9 μs
MLX (Apple Silicon) 3.12M/sec 4.27M/sec 5.95M/sec 1.08 μs

*MLX achieves 7-40x throughput improvement via Metal GPU parallelism

0.2.0 - 2025-01-19

Added

  • External MLX package integration (github.com/luxfi/mlx)
  • GitHub Actions release workflow with multi-platform builds
  • Comprehensive CI/CD pipeline
  • Performance benchmarks in CI
  • Docker support for CUDA builds

Changed

  • Replaced local MLX implementations with external package
  • Updated CI workflows for better test coverage
  • Improved documentation structure
  • Simplified build system

Removed

  • Redundant bridge/ directory (CGO bridge now in luxfi/mlx)
  • Local replace directives from go.mod
  • Duplicate MLX implementations
  • Unnecessary documentation files

Performance

  • Order matching: 1181 ns/op (under 2μs target)
  • Concurrent operations: 1738 ns/op
  • Throughput: 847K orders/sec (CPU)

0.1.0 - 2025-01-18

Added

  • Initial release
  • Ultra-high performance order book
  • Multi-engine architecture (Go, C++, MLX)
  • Quantum-resistant consensus (FPC)
  • Basic benchmarking suite
  • Demo application

Performance

  • Achieved 434M+ orders/sec with MLX GPU
  • Sub-microsecond latency for order matching

0.0.1 - 2025-01-15

Added

  • Project initialization
  • Basic order book implementation
  • Test suite framework