rtx 3060 notebook cuda 与 gpu驱动安装

cudnn的下载与安装：

需要用户登录nvidia，然后再进入该地址：

https://developer.nvidia.com/rdp/cudnn-archive

https://developer.download.nvidia.cn/compute/cuda/opensource/

https://developer.nvidia.com/gpu-accelerated-libraries

CUPTI安装：

https://docs.nvidia.com/cupti/Cupti/index.html

cuda11.2.2中，CUPTI是随着cuda一起安装的，不过路径有变动，需要手动copy到对应目录：

比如，复制如下文件：

D:\Program Files\NVIDIAGPUComputingToolkit\CUDA\v11.2\extras\CUPTI\lib64\cupti64_2020.3.1.dll

粘贴至如下目录：

D:\Program Files\NVIDIAGPUComputingToolkit\CUDA\v11.2\bin\

cudnn安装在win10上，指导书：

https://docs.nvidia.com/deeplearning/cudnn/install-guide/index.html#install-windows

下载、解压、拷贝粘贴其中的dll、inlucde头文件和lib，至相应的目录。如文中所言：

1. Copy the following files into the CUDA Toolkit directory.
   1. Copy <installpath>\cuda\bin\cudnn*.dll to C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\vx.x\bin.
   2. Copy <installpath>\cuda\include\cudnn*.h to C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\vx.x\include.
   3. Copy <installpath>\cuda\lib\x64\cudnn*.lib to C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\vx.x\lib\x64.

TensorRT下载地址：

https://developer.nvidia.com/nvidia-tensorrt-download