Skip to content

v2.0.0 - The Unmanaged Arena Release

Choose a tag to compare

@MachineGate MachineGate released this 01 Aug 09:17
· 11 commits to main since this release

🚀 Release v2.0.0 (Initial Public Release)

Welcome to the first public release of MachineGate.NG.Velox! This is a complete systems-level rewrite optimized for .NET 10, delivering a zero-allocation, ultra-high-performance G-code processing pipeline with native C/C++ speed.


⚡ Performance & Core Architecture

  • Absolute Zero-Allocation (GC-Free): The entire end-to-end pipeline operates with 0 B managed heap allocations using a custom unmanaged MemoryArena. Garbage Collection pauses are completely eliminated.
  • Extreme Throughput: Processes ~400,000 lines of G-code per second (validated on Intel i5-10400F, .NET 10.0, X64 RyuJIT AVX2).
  • Systems-Level Optimizations:
    • Branchless lexing via CharRegistry expanded ushort bitmasks.
    • Explicit 16-byte AST Node struct unions for maximum L1/L2 cache locality.
    • Raw pointer arithmetic (T*) in hot paths to bypass array bounds checking.

🛠 Industrial Kinematics & Dual Pipeline

  • Advanced Motion Planning: Implements full S-Curve (jerk-limited) acceleration profiles, Look-Ahead planning, and Junction Deviation.
  • UVW Plane Normalization: Unified mathematical paths for G17/G18/G19 planes.
  • Dual Output Targets:
    • Hardware Target: Packs toolpaths into dense, zero-copy binary structures for direct UART/Ethernet microcontroller transmission.
    • Simulation Target: Generates unmanaged SimulationFrame structures pointing directly to arena memory for high-speed visualization.
  • Precise Diagnostics: Integrated IndexMap to trace preprocessed characters back to original source code lines.

📦 Installation

dotnet add package MachineGate.NG.Velox --version 2.0.0

Note: For complete implementation details and code snippets for both Hardware and Simulation pipelines, please refer to the updated README.md in the repository root.