适用于_Linux_Debian_12_安装Tesla_P4_P40等GPU的显卡Cuda和驱动_

# Nvidia_Tesla_P4_P40_Manual_For_Debian_12 适用于 Linux Debian 12 安装Tesla P4 P40等GPU的显卡Cuda和驱动 Suitable for Linux Debian 12 to install graphics card Cuda and driver for Tesla P4 P40 and other GPUs 在pve 8.2中使用q35（不可使用i440fx)机型创建的虚拟机中安装Nvidia gpu驱动的教程。 Tutorial for installing Nvidia gpu drivers in a virtual machine created with the q35 (not available with i440fx) model in pve 8.2.   ## 检查显卡存在 通常可以使用 lspci 命令来识别已安装显卡的 NVIDIA 图形处理单元 (GPU) 系列/代号。例如： You can usually use the lspci command to identify the NVIDIA graphics processing unit (GPU) family/codename of an installed graphics card. For example: ```bash $ lspci | grep NVIDIA ``` ## 为apt允许非自由软件源 Allow non-free software sources for apt ```bash for non-Chinese users: vim.tiny /etc/apt/sources.list Add "contrib", "non-free" and "non-free-firmware" components to /etc/apt/sources.list, for example: # Debian Bookworm deb http://deb.debian.org/debian/ bookworm main contrib non-free non-free-firmware ``` ```bash # 对于中国用户更换清华tuna源： vim.tiny /etc/apt/sources.list deb https://mirrors.tuna.tsinghua.edu.cn/debian/ bookworm main contrib non-free non-free-firmware # deb-src https://mirrors.tuna.tsinghua.edu.cn/debian/ bookworm main contrib non-free non-free-firmware deb https://mirrors.tuna.tsinghua.edu.cn/debian/ bookworm-updates main contrib non-free non-free-firmware # deb-src https://mirrors.tuna.tsinghua.edu.cn/debian/ bookworm-updates main contrib non-free non-free-firmware deb https://mirrors.tuna.tsinghua.edu.cn/debian/ bookworm-backports main contrib non-free non-free-firmware # deb-src https://mirrors.tuna.tsinghua.edu.cn/debian/ bookworm-backports main contrib non-free non-free-firmware deb https://mirrors.tuna.tsinghua.edu.cn/debian-security bookworm-security main contrib non-free non-free-firmware # deb-src https://mirrors.tuna.tsinghua.edu.cn/debian-security bookworm-security main contrib non-free non-free-firmware ``` ### 更新apt ### Update apt ```bash apt update -y ``` ## 安装显卡驱动 ```bash apt install nvidia-driver firmware-misc-nonfree ``` 重启电脑 ### 检查 Debian 12 上是否安装了显卡驱动 ```bash nvidia-smi +---------------------------------------------------------------------------------------+ | NVIDIA-SMI | |-----------------------------------------+----------------------+----------------------+ | GPU Name Persistence-M | Bus-Id Disp.A | Volatile Uncorr. ECC | | Fan Temp Perf Pwr:Usage/Cap | Memory-Usage | GPU-Util Compute M. | | | | MIG M. | |=========================================+======================+======================| | 0 Tesla P4 On | 00000000:03:00.0 Off | Off | | N/A 38C P8 6W / 75W | 0MiB / 8192MiB | 0% Default | | | | N/A | +-----------------------------------------+----------------------+----------------------+ +---------------------------------------------------------------------------------------+ | Processes: | | GPU GI CI PID Type Process name GPU Memory | | ID ID Usage | |=======================================================================================| | No running processes found | +---------------------------------------------------------------------------------------+ ``` 显卡驱动已安装完毕 ## 安装Cuda install Cuda ```bash apt install nvidia-cuda-dev nvidia-cuda-toolkit ``` ### 检查 Debian 12 上是否安装了 NVIDIA CUDA ```bash nvcc --version ``` ## 安装 NVIDIA cuDNN install cuDNN ```bash apt install nvidia-cudnn ``` 看到窗口后，按 <Enter> 键同意。 Once you see the window, press <Enter> to agree the license. NVIDIA cuDNN 库需要从 NVIDIA 官方网站下载。需要一段时间。 NVIDIA cuDNN libraries are being downloaded from the official website of NVIDIA. It takes a while to complete. ## 关闭ECC Turn off ECC 通过```nvidia-smi | grep Tesla```查看前面GPU编号 ```bash d@d:/fuck$ nvidia-smi | grep Tesla | 0 Tesla P40 On | 00000000:03:00.0 Off | Off | ----------------------------------------------------------------------------------------- nvidia-smi -i n -e 0/1 可关闭(0)/开启(1) , n是GPU的编号。 ``` 执行关闭ECC```sudo nvidia-smi -i 0 -e 0```重启后该设置生效。