Skip to content

README.md

Latest commit

 

History

History
74 lines (53 loc) · 2.31 KB

README.md

File metadata and controls

74 lines (53 loc) · 2.31 KB

vector_math

A C++20 vector and matrix mathematics library with SIMD acceleration for x86/x64 (SSE/AVX) and ARM (NEON).

🚀 Part of the Campello Engine

This project is a module within the Campello ecosystem.

👉 Main repository: https://github.com/rusoleal/campello

Campello is a modular, composable game engine built as a collection of independent libraries. Each module is designed to work standalone, but integrates seamlessly into the engine runtime.

Features

  • Generic vector and matrix templates (Vec<T, N>, Mat<T, R, C>)
  • Typed specializations: Vector2, Vector3, Vector4, Matrix2, Matrix3, Matrix4
  • SIMD-optimized float paths by default on x86/x64 (SSE) and optional AVX2-accelerated double paths on x86/x64
  • Quaternion<T> with Hamilton product, slerp, axis-angle, and vector rotation
  • Rich vector API: dot, cross, distance, angle, reflect, lerp, clamp, floor/ceil/round, component-wise ops
  • Rich matrix API: determinant, inverse, TRS compose/decompose helpers, row/column access
  • Full Doxygen documentation on all public APIs
  • Namespace: systems::leal::vector_math

Requirements

  • CMake 3.22.1+
  • C++20 compiler (GCC, Clang, MSVC)

Building

cmake -B build -S .
make -C build

Enable AVX2 explicitly on x86/x64 when you want the wider double-precision SIMD paths:

cmake -B build -S . -DVECTOR_MATH_ENABLE_AVX2=ON
make -C build

Tests

./launch_test.sh
# or
cmake -B build/test -S . -DVECTOR_MATH_BUILD_TEST=ON
make -C build/test
build/test/vector_math_test

Benchmarks

./launch_benchmark.sh
# or
cmake -B build/benchmark -S . -DVECTOR_MATH_BUILD_BENCHMARK=ON -DCMAKE_BUILD_TYPE=Release
make -C build/benchmark
build/benchmark/vector_math_benchmark

SIMD Support

Architecture is detected automatically at compile time:

Architecture Intrinsics Types accelerated
x86/x64 SSE by default, AVX2+FMA optional via -DVECTOR_MATH_ENABLE_AVX2=ON Vector4f, Vector3f, Matrix4f by default; Vector4d, Vector3d, Matrix4d use AVX when enabled
AArch64 NEON Matrix4f (NEON), Matrix4d (NEON)
ARM 32-bit — Matrix4f (scalar fallback), Matrix4d (scalar fallback)

Vector3f/Vector4f/Matrix4f use alignas(16). Vector3d/Vector4d/Matrix4d use alignas(32) for AVX-capable layouts.