我要评分
获取效率
正确性
完整性
易理解

NEON Library Application Acceleration

Principles

Some typical libraries developed based on the NEON technology are used to call function module interfaces to accelerate applications. In this way, developers can obtain the maximum benefits without using complicated NEON assembly or NEON intrinsic functions for code acceleration.

Modification Method

  • Arm Compute Library (ACL)

    Arm Compute Library is an open source project released by Arm. It aims to provide hardware acceleration libraries of the Arm platform for developers in the image, video, multimedia, and computer vision fields. In this library, OpenCL and NEON are used to implement some basic algorithms in the preceding fields. OpenCL is used for Mali GPU acceleration of Arm, and NEON is used for the A-series CPUs of Arm. The library contains function modules such as basic mathematical matrix operations, basic image processing operations, and basic machine learning algorithms.

    Reference link: https://developer.arm.com/technologies/compute-library

    GitHub address: https://github.com/ARM-software/ComputeLibrary

  • Ne10

    Ne10 is an open source C library that is implemented based on the NEON assembly of C interfaces. It is hosted by Arm on GitHub and contains a group of common processing-intensive functions. These functions are optimized based on Arm. Ne10 adopts the modular structure and consists of several small libraries. It includes math functions, signal processing functions, image processing functions and physics functions.

    Reference link: https://github.com/projectNe10/Ne10

  • libyuv

    libyuv is an open-source basic library that implements conversion, rotation, and scaling of YUV and RGB images. It is developed by Google based on the SIMD instruction set.

    Reference link: https://code.google.com/p/libyuv/

  • Skia

    Skia is an open-source 2D graphics library for graphics engines in Google Chrome, Chrome OS, Android, Mozilla Firefox, Firefox OS, and many other products.

    Reference link: https://github.com/google/skia