1. 程式人生 > >深度學習開發環境配置第一彈:Ubuntu16.04下安裝NVIDIA顯示卡驅動+CUDA9.0.176配置

深度學習開發環境配置第一彈:Ubuntu16.04下安裝NVIDIA顯示卡驅動+CUDA9.0.176配置

一、解除安裝舊NVIDIA驅動

sudo apt-get remove --purge nvidia*

二、拉黑Ubuntu核心裡面自帶的nouveau驅動

本人親測使用的是:

1、在/etc/modprobe.d/路徑新增並修改配置檔案

[email protected]:~$ sudo touch /etc/modprobe.d/blacklist-nouveau.conf
[email protected]:~$ sudo nano /etc/modprobe.d/blacklist-nouveau.conf

在blacklist-nouveau.conf檔案中新增這幾句話:

blacklist nouveau

options nouveau modeset=0

2、 重新生成 kernel initramfs

[email protected]:~$ sudo update-initramfs -u
update-initramfs: Generating /boot/initrd.img-4.4.0-21-generic
W: Possible missing firmware /lib/firmware/ast_dp501_fw.bin for module ast

3、重啟

[email protected]:~$ sudo reboot

4、最終確認nouveau是否被拉黑

輸入lsmod | grep nouveau檢視,沒有輸出即禁止成功

拉黑前:

[email protected]:~$ lsmod | grep nouveau
nouveau              1495040  0
mxm_wmi                16384  1 nouveau
video                  40960  1 nouveau
ttm                    98304  2 ast,nouveau
drm_kms_helper        139264  2 ast,nouveau
i2c_algo_bit           16384  3 ast,igb,nouveau
drm                   360448  6 ast,ttm,drm_kms_helper,nouveau
wmi                    20480  2 mxm_wmi,nouveau

 拉黑後:

[email protected]:~$ lsmod | grep nouveau
[email protected]:~$ sudo service lightdm stop

三、安裝NVIDIA顯示卡驅動

工作站reboot後analt + ctrl + F1進入textmode模式,輸入使用者和密碼後,禁用lightdm服務:

[email protected]:~$ sudo service lightdm stop

進入NVIDIA顯示卡驅動官網下載驅動(這種下載方式貌似只提供最新的驅動版本下載),切換到驅動.run檔案路徑,使用sudo ./NVIDIA*.run -no-opengl-files安裝NVIDIA驅動(後面的-no-opengl-files不能省掉,否則會造成進不去桌面,卡在登陸介面迴圈):

[email protected]:/media/vslyu/home/sinc-lab$ sudo ./Documents/ubantu伺服器軟體/NVIDIA-Linux-x86_64-390.59.run -no-opengl-files
Verifying archive integrity... OK
Uncompressing NVIDIA Accelerated Graphics Driver for Linux-x86_64 390.59........................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................

安裝完成後,輸入nvidia-smi命令檢視顯示卡資訊是否存在,存在即說明成功打上驅動:

[email protected]:/media/vslyu/home/sinc-lab/Downloads$ nvidia-smi
Tue Sep 11 17:23:40 2018
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 384.130                Driver Version: 384.130                   |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|===============================+======================+======================|
|   0  GeForce GTX 108...  Off  | 00000000:02:00.0 Off |                  N/A |
| 23%   41C    P0    62W / 250W |      0MiB / 11172MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+
|   1  GeForce GTX 108...  Off  | 00000000:03:00.0 Off |                  N/A |
| 24%   44C    P0    61W / 250W |      0MiB / 11172MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+
|   2  GeForce GTX 108...  Off  | 00000000:82:00.0 Off |                  N/A |
| 23%   41C    P0    61W / 250W |      0MiB / 11172MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+
|   3  GeForce GTX 108...  Off  | 00000000:83:00.0 Off |                  N/A |
| 23%   39C    P0    59W / 250W |      0MiB / 11172MiB |      2%      Default |
+-------------------------------+----------------------+----------------------+

+-----------------------------------------------------------------------------+
| Processes:                                                       GPU Memory |
|  GPU       PID   Type   Process name                             Usage      |
|=============================================================================|
|  No running processes found                                                 |
+-----------------------------------------------------------------------------+

(ps:之前安裝的Driver Version為390.87,單這裡輸出的 Driver Version為384.130,這是因為我這裡貼的圖是安裝自帶NVIDIA Driver的本地cuda-repo-ubuntu1604-9-0-local_9.0.176-1_amd64.deb檔案後的輸出結果)

四、安裝CUDA

參照NVIDIA的NVIDIA CUDA Installation Guide for Linux,按照要求一步一步進行確認,推薦使用.deb檔案進行安裝,使用.run檔案進行安裝時之前在配置TensorRT遇到過一個坑(不好意思,這個坑具體是什麼我忘了),

線上安裝的deb(安裝線上的deb會提醒選擇Driver Version的是否安裝、是否需要更新等資訊):

sudo dpkg -i cuda-repo-ubuntu1604_9.0.176-1_amd64.deb
~$ sudo apt-key adv --fetch-keys http://developer.download.nvidia.com/compute/cuda/repos/ubuntu1604/x86_64/7fa2af80.pub
~$ sudo apt-get update
~$ sudo apt-get install cuda

本地安裝的deb(安裝線上的deb不會提醒選擇Driver Version的是否安裝、是否需要更新等資訊,預設配置本地自帶的Driver,故安裝本地的deb檔案需要先配置Driver,安裝本地的deb檔案會自動安裝配置好):

[email protected]:/media/vslyu/home/sinc-lab/Downloads$ sudo dpkg -i cuda-repo-ubuntu1604-9-0-local_9.0.176-1_amd64.deb
~$ sudo apt-key adv --fetch-keys http://developer.download.nvidia.com/compute/cuda/repos/ubuntu1604/x86_64/7fa2af80.pub
~$ sudo apt-get update
~$ sudo apt-get install cuda

注意新增.pub公鑰後一定要update一下,建立好apt-cache資訊:

[email protected]:/media/vslyu/home/sinc-lab/Downloads$ sudo apt-get update
Get:1 file:/var/cuda-repo-9-0-local  InRelease
Ign:1 file:/var/cuda-repo-9-0-local  InRelease
Get:2 file:/var/cuda-repo-9-0-local  Release [574 B]
Hit:3 http://mirrors.hust.edu.cn/ubuntu xenial InRelease
Get:2 file:/var/cuda-repo-9-0-local  Release [574 B]
Get:4 http://mirrors.hust.edu.cn/ubuntu xenial-security InRelease [107 kB]
Get:5 http://mirrors.hust.edu.cn/ubuntu xenial-updates InRelease [109 kB]
Hit:6 http://mirrors.hust.edu.cn/ubuntu xenial-proposed InRelease
Get:7 http://mirrors.hust.edu.cn/ubuntu xenial-backports InRelease [107 kB]
Get:9 http://mirrors.hust.edu.cn/ubuntu xenial-security/universe Sources [73.6 kB]
Get:10 http://mirrors.hust.edu.cn/ubuntu xenial-security/main amd64 DEP-11 Metadata [67.7 kB]
Get:11 http://mirrors.hust.edu.cn/ubuntu xenial-security/main DEP-11 64x64 Icons [68.0 kB]
Get:12 http://mirrors.hust.edu.cn/ubuntu xenial-security/universe amd64 Packages [370 kB]
Get:13 http://mirrors.hust.edu.cn/ubuntu xenial-security/universe amd64 DEP-11 Metadata [107 kB]
Get:14 http://mirrors.hust.edu.cn/ubuntu xenial-security/universe DEP-11 64x64 Icons [142 kB]
Get:15 http://mirrors.hust.edu.cn/ubuntu xenial-updates/universe Sources [220 kB]
Get:16 http://mirrors.hust.edu.cn/ubuntu xenial-updates/main amd64 DEP-11 Metadata [320 kB]
Get:17 http://mirrors.hust.edu.cn/ubuntu xenial-updates/main DEP-11 64x64 Icons [227 kB]
Get:18 http://mirrors.hust.edu.cn/ubuntu xenial-updates/universe amd64 Packages [683 kB]
Get:19 http://mirrors.hust.edu.cn/ubuntu xenial-updates/universe i386 Packages [624 kB]
Get:20 http://mirrors.hust.edu.cn/ubuntu xenial-updates/universe amd64 DEP-11 Metadata [247 kB]
Get:21 http://mirrors.hust.edu.cn/ubuntu xenial-updates/universe DEP-11 64x64 Icons [335 kB]
Get:22 http://mirrors.hust.edu.cn/ubuntu xenial-updates/multiverse amd64 DEP-11 Metadata [5,960 B]
Get:23 http://mirrors.hust.edu.cn/ubuntu xenial-updates/multiverse DEP-11 64x64 Icons [14.3 kB]
Get:24 http://mirrors.hust.edu.cn/ubuntu xenial-backports/main amd64 DEP-11 Metadata [3,328 B]
Get:25 http://mirrors.hust.edu.cn/ubuntu xenial-backports/universe amd64 DEP-11 Metadata [5,100 B]
Fetched 3,835 kB in 0s (4,534 kB/s)
Reading package lists... Done
[email protected]:/media/vslyu/home/sinc-lab/Downloads$

在環境變數中新增CUDA,sudo許可權開啟~/.bashrc檔案,在檔案末尾新增兩行指令:

export PATH=/usr/local/cuda-9.0/bin${PATH:+:${PATH}}
export LD_LIBRARY_PATH=/usr/local/cuda-9.0/lib64\ ${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}

再啟用、載入一下配置:

[email protected]:/media/vslyu/home/sinc-lab/Downloads$  source ~/.bashrc
[email protected]:/media/vslyu/home/sinc-lab/Downloads$ sudo ldconfig
[email protected]:/media/vslyu/home/sinc-lab/Downloads$