一种基于ARMv8架构CPU的算法加速方法

A method of accelerating algorithms based on CPU with ARMv8 architecture

基于目前主流的硬件平台ARMv8架构CPU,提出一种使用SIMD技术的算法加速方法,以atan2函数为例,与标准库的性能作对比。实验结果表明,在FT1500计算平台上,采用此种方式优化的算法性能有显著提升。

Based on the current mainstream hardware platform,the CPU with the ARMv8 architecture,a method of accelerating algorithms is proposed by using the single instruction multiple data(SIMD)technology.Taking atan2 function as an example,compared with that of the standard library,the performance of the algorithms optimized in this way is significantly improved on the FT1500 computing platform.