本文主要记录如何在device代码内异步把数据从全局内存复制至shared内存，有关如何异步把数据从主机端拷贝到设备端，可以参考How to Overlap Data Transfers in CUDA C/C++ | NVIDIA Technical Blog 本文主要内容来源于英伟达博客：Controlling Data Movement to Boost Performance on...