Ubuntu20.04LTS+Cuda11.3安装 驱动

Ubuntu20.04LTS+Cuda11.3安装 驱动

解决驱动 循环登录 报错

设备戴尔G7-7590

Intel® Core™ i7-9750H CPU @ 2.60GHz × 12 + GeForce RTX 2060 双显卡

Win10+Ubuntu20.04的双系统

我选择显卡驱动和cuda一起安装

• 查看英伟达显卡,会显示NVIDIA GPU 的信息

  lspci | grep -I nvidia  
  lspci | grep -i nvidia
  

• 查看自己的linux

  uname -m && cat /etc/*release
  

• 查看正在运行的系统内核版本，安装对应的kernel header和package development

  uname -r
  sudo apt-get install linux-headers-$(uname -r)
  

• remove之前的驱动

  sudo apt-get remove --purge nvidia*
  

• 安装gcc,g++,make等工具

  sudo apt-get update
  sudo apt-get install dkms build-essential   linux-headers-generic
  

• 禁用nouveau

  sudo apt-get install vim
  sudo vim /etc/modprobe.d/blacklist.conf
  

• 在最后加上

  blacklist nouveau
  blacklist lbm-nouveau
  options nouveau modeset=0
  alias nouveau off
  alias lbm-nouveau off
  

• 禁用nouveau内核

  echo options nouveau modeset=0
  sudo update-initramfs -u
  

• 重启电脑，运行下面指令，如果nouveau禁用成功，则无输出。

  lsmod | grep nouveau
  

• 下载cuda11.3（官网命令runfile）

  wget https://developer.download.nvidia.com/compute/cuda/11.3.1/local_installers/cuda_11.3.1_465.19.01_linux.run
  

• 关闭图形界面并且安装cuda

  sudo systemctl set-default multi-user.target
  sudo sh cuda_11.3.1_465.19.01_linux.run
  reboot
  

• 重启后应该进不去桌面，ctrl atl +F2或者F4等进入命令行模式，log in，开启图形界面，reboot

  sudo systemctl set-default graphical.target
  sudo reboot
  

• 添加环境变量

  sudo vim ~/.bashrc
  

• 最后一行加入（vim用i进入insert）

  export CUDA_HOME=/usr/local/cuda
  export PATH=$PATH:$CUDA_HOME/bin
  export LD_LIBRARY_PATH=/usr/local/cuda-11.3/lib64${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}
  

• 使其生效

  source ~/.bashrc
  

• 检查命令：

  cat /proc/driver/nvidia/version
  nvcc -V
  nvidia-smi
  cd /usr/local/cuda/samples/1_Utilities/deviceQuery
  sudo make
  ./deviceQuery
  

• 第一个cuda代码（guide上查看设备性能的code）

  vim Device.cu
  

  #include <stdio.h>
  int main() {
          int nDevices;
  
          cudaGetDeviceCount(&nDevices);
          for (int i = 0; i < nDevices; i++) {
                  cudaDeviceProp prop;
                  cudaGetDeviceProperties(&prop, i);
                  printf("Device Num: %d\n", i);
                  printf("Device name: %s\n", prop.name);
                  printf("Device SM Num: %d\n", prop. multiProcessorCount);
                  printf("Share Mem Per Block: %.2fKB\n", prop.   sharedMemPerBlock / 1024.0);
                  printf("Max Thread Per Block: %d\n", prop.  maxThreadsPerBlock);
                  printf("Memory Clock Rate (KHz): %d\n",
                     prop.memoryClockRate);
                  printf("Memory Bus Width (bits): %d\n",
                     prop.memoryBusWidth);
                  printf("Peak Memory Bandwidth (GB/s): %.    2f\n\n",
                     2.0 * prop.memoryClockRate * (prop.  memoryBusWidth / 8) / 1.0e6);
          }
          return 0;
  }
  

• 编译并且运行

  nvcc Device.cu -o Device.o
  ./Device.o
  

• 解决nvvp无法运行的问题——安装java

  sudo apt install openjdk-8-jdk
  

• To run Visual Profiler on Ubuntu

  Make sure that you invoke Visual Profiler with the command-line option included as shown below:

  nvvp -vm /usr/lib/jvm/java-8-openjdk-amd64/jre/bin/java
  

  Note: The-vmoption is only required when JRE is not included in CUDA Toolkit package and JRE 1.8 is not in the default path.

  vim ~/.bashrc
  alias nvvp='nvvp -vm /usr/lib/jvm/java-8-openjdk-amd64/jre/ bin/java'
  source ~/.bashrc