- A C library that performs matrix/vector algebraic operations with NEON ARM intrinsic
- The Project is very similar to Ne10 project by ARM and contributors(All credits to Ne10 developers for giving such basic and well developed source for ARM intrinsic).
- But Ne10 is not supported for Windows on ARM, Since the project is originally written for Unix based ARM devices.
- MATHNEON mainly focuses on supporting optimized math routines on Windows on ARM devices and help users/developers providing enough knowledge towards NEON intrinsic in a very basic way.
- Currently, MATHNEON supports Int32 and Float32 math rountines and the development is still in progress.
- Contributions and ideas are always welcome, please feel free to open an issue to start discussion.
- As a beginner, I always recommend people to start with understanding SIMD - https://en.wikipedia.org/wiki/Single_instruction,_multiple_data
- If you get enough hold SIMD, start with understanding NEON SIMD - https://developer.arm.com/Architectures/Neon
- Ne10 for basics and understanding MATHNEON code base - https://github.com/projectNe10/Ne10/tree/master
- List of ARM NEON intrinsic used in MATHNEON code base - https://arm-software.github.io/acle/neon_intrinsics/advsimd.html