MathLib

MathLib is a cross-platform header-only SSE/AVX/NEON-accelerated math library, designed for computer graphics.

Features:

compile-time optimization level specialization: SSE3, SSE4, AVX1, AVX2 (or NEON on ARM via sse2neon)
float3, float4 and float4x4 types
double3, double4 and double4x4 types
int2, uint2, int4 and uint4 types
common functions: abs, round, floor, mod, clamp, saturate, min, max, and more...
transcendental functions: sin, cos, tan, sqrt, rsqrt, asin, acos, and more...
data conversion and packing functionality - FP32, FP16, SNORM and UNORM (with any number of bits per component)
vectors and matrices
linear algebra miscellaneous functionality
projective math miscellaneous functionality
frustum & AABB primitives
random numbers generation
sorting

Also includes STL.hlsli file which is a standalone HLSL math library.

License

MathLib is licensed under the MIT License.