ubuntu 同时安装cuda8.0与cuda9.0,cuda9.1

部分代码需要cuda8.0，部分需要cuda9.0 于是萌生了同时安装2个版本的想法。

0 前提：

ubuntu 16.04. x86_64

已经安装 cuda 8.0:

nvcc –version
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2016 NVIDIA Corporation
Built on Tue_Jan_10_13:22:03_CST_2017
Cuda compilation tools, release 8.0, V8.0.61

1 下载 cuda 9.0：

cuda 9.0

2.  在自己的目录下为cuda9.0新建一个文件夹，用于存放  cuda_9.0.176_384.81_linux.run 与后面的sample。

3.  生成可执行文件 

 chmod 777 cuda_9.0.176_384.81_linux.run

4. 执行：

./cuda_9.0.176_384.81_linux.run

5. 一些安装过程：

(因为我需要保留cuda8.0 于是暂时不做软连接 )

Do you want to install a symbolic link at /usr/local/cuda?
(y)es/(n)o/(q)uit: n 

Do you accept the previously read EULA?
accept/decline/quit: accept

Install NVIDIA Accelerated Graphics Driver for Linux-x86_64 384.81?
(y)es/(n)o/(q)uit: y

Do you wish to run the installation with 'sudo'?
(y)es/(n)o: y

Please enter your password: 
Do you want to install the OpenGL libraries?
(y)es/(n)o/(q)uit [ default is yes ]: 

Do you want to run nvidia-xconfig?
This will update the system X configuration file so that the NVIDIA X driver
is used. The pre-existing X configuration file will be backed up.
This option should not be used on systems that require a custom
X configuration, such as systems with multiple GPU vendors.
(y)es/(n)o/(q)uit [ default is no ]:  

Install the CUDA 9.0 Toolkit?
(y)es/(n)o/(q)uit: y

Enter Toolkit Location
 [ default is /usr/local/cuda-9.0 ]: 

Do you want to install a symbolic link at /usr/local/cuda?
(y)es/(n)o/(q)uit: n 

Install the CUDA 9.0 Samples?
(y)es/(n)o/(q)uit: y

Enter CUDA Samples Location
 [ default is /home/linlf ]: /home/linlf/tools/cuda9.0

5报错：

Itappears that an X server is running. Please exit X beforeinstallation. If you're sure that X is not running, but are gettingthis error, please delete any X lock files in /tmp.

6 参考 https://blog.csdn.net/cc1949/article/details/78491781

6.1 查看 X：psaux | grep X

6.2 关闭X. light类型的桌面系统

      Sudo /etc/init.d/lightdm stop

6.4 接着按照第一步第方法继续安装cuda9.0

7  修改配置文件：

export PATH="/usr/local/cuda/bin:$PATH"
export LD_LIBRARY_PATH="/usr/local/cuda/lib64:$LD_LIBRARY_PATH"
export LIBRARY_PATH="/usr/local/cuda/lib64:$LIBRARY_PATH"

8。 配置文件生效：

source ~/.bashrc

9。然后就可以切换8.0与 9.0了：

rm –rf /usr/local/cuda

ln -s /usr/local/cuda-9.0 /usr/local/cuda

10。若别的用户需要使用到。同样如上切换。

11。 最后进入 cuda9.0的sample。make所有第例子.

(后续：)

12 若需要再安装cuda9.1: 会报错：

 15 Using built-in stream user interface
 16 -> Detected 56 CPUs online; setting concurrency level to 32.
 17 -> License accepted by command line option.
 18 -> Installing NVIDIA driver version 387.26.
 19 -> The NVIDIA driver appears to have been installed previously using a different installer. To prevent potential conflicts, it is recommended either to update the
     existing installation using the same mechanism by which it was originally installed, or to uninstall the existing installation before installing this driver. 20 
 21 Please review the message provided by the maintainer of this alternate installation method and decide how to proceed:
 22 
 23 The package that is already installed is named nvidia-384.
 24 
 25 You can upgrade the driver by running:
 26 `apt-get install nvidia-384 nvidia-modprobe nvidia-settings`
 27 
 28 You can remove nvidia-384, and all related packages, by running:
 29 `apt-get remove --purge nvidia-384 nvidia-modprobe nvidia-settings`
 30 
 31 This package is maintained by NVIDIA (cudatools@nvidia.com).
 32 
 33 
 34 (Answer: Abort installation)
 35 ERROR: The installation was canceled due to the availability or presence of an alternate driver installation. Please see /var/log/nvidia-installer.log for more de
    tails.

意思是CUDA    驱动版本冲突：

CUDA8.0依赖的是384版本驱动，而CUDA9.1依赖的是387版本驱动，因此需要更新/卸载384驱动。

卸载：

sudo apt-get remove --purge nvidia-384 nvidia-modprobe nvidia-settings

卸载细节：

lin@cc:~/tools/cuda9.1$ sudo apt-get remove --purge nvidia-384 nvidia-modprobe n
vidia-settingsReading package lists... Done
Building dependency tree       
Reading state information... Done
The following packages were automatically installed and are no longer required:
  apt-xapian-index aptitude-common cuda-command-line-tools-8-0 cuda-core-8-0
  cuda-cublas-8-0 cuda-cublas-dev-8-0 cuda-cudart-8-0 cuda-cudart-dev-8-0
  cuda-cufft-8-0 cuda-cufft-dev-8-0 cuda-curand-8-0 cuda-curand-dev-8-0
  cuda-cusolver-8-0 cuda-cusolver-dev-8-0 cuda-cusparse-8-0 cuda-cusparse-dev-8-0
  cuda-documentation-8-0 cuda-driver-dev-8-0 cuda-license-8-0 cuda-misc-headers-8-0
  cuda-npp-8-0 cuda-npp-dev-8-0 cuda-nvgraph-8-0 cuda-nvgraph-dev-8-0 cuda-nvml-dev-8-0
  cuda-nvrtc-8-0 cuda-nvrtc-dev-8-0 cuda-samples-8-0 cuda-toolkit-8-0
  cuda-visual-tools-8-0 dh-apparmor fonts-dejavu g++-4.8 g++-4.8-multilib
  gir1.2-gconf-2.0 gir1.2-gnomebluetooth-1.0 gstreamer1.0-clutter iproute
  lib32stdc++-4.8-dev libatk-wrapper-java libatk-wrapper-java-jni libavcodec54
  libavformat54 libavutil52 libboost-atomic1.54.0 libboost-chrono1.54.0
  libboost-context1.54.0 libboost-date-time1.54.0 libboost-filesystem1.54.0
  libboost-graph1.54.0 libboost-iostreams1.54.0 libboost-locale1.54.0
  libboost-log1.54.0 libboost-math1.54.0 libboost-program-options1.54.0
  libboost-python1.54.0 libboost-random1.54.0 libboost-regex1.54.0
  libboost-serialization1.54.0 libboost-signals1.54.0 libboost-system1.54.0
  libboost-test1.54.0 libboost-thread1.54.0 libboost-timer1.54.0 libboost-wave1.54.0
  libc-ares2 libcamel-1.2-45 libclutter-gst-2.0-0 libcolord1 libcolorhug1 libcr0
  libcwidget3 libdbi1 libebackend-1.2-7 libebook-1.2-14 libebook-contacts-1.2-0
  libedata-book-1.2-20 libedataserver-1.2-18 libegl1-mesa-drivers libelfg0 libept1.4.12
  libgcrypt11-dev libgdata13 libgee2 libgif4 libgnome-bluetooth11 libgnome-desktop-3-7
  libgnutlsxx27 libgphoto2-port10 libgtop2-7 libicu52 libilmbase6 libimobiledevice4
  libisl10 libjavascriptcoregtk-3.0-0 libllvm3.4 libmbim-glib0 libmysqlclient-dev
  libmysqlclient18 libopenexr6 libopenjpeg2 libopenvg1-mesa libpango1.0-0 libpci-dev
  libplist1 libpoppler44 libqmi-glib0 libqpdf13 libqt5positioning5 libqt5sensors5
  libqt5test5 librrd4 libservlet3.0-java libsigc++-2.0-0c2a libsqlite3-dev libswscale2
  libsystemd-journal0 libtorque2 libupower-glib1 libusbmuxd2 libv8-3.14.5 libvpx1
  libvte-2.90-9 libvte-2.90-common libwebkitgtk-3.0-0 libwebkitgtk-3.0-common
  libx264-142 libx32stdc++-4.8-dev libxcb-util0 libxen-4.4 libzeitgeist-1.0-1
  linux-image-3.13.0-145-generic linux-image-extra-3.13.0-145-generic
  linux-image-generic-lts-xenial linux-lts-xenial-cloud-tools-4.4.0-31
  linux-lts-xenial-tools-4.4.0-31 nvidia-prime obex-data-server php5-cli php5-common
  php5-json php5-readline python-colorama python-cups python-dbus-dev python-distlib
  python-gnomekeyring python-gobject python-html5lib python-notify python-pycurl
  python-smbc python-urlgrabber python-xapian python3-colorama python3-distlib
  python3-imaging ttf-dejavu ttf-dejavu-core ttf-dejavu-extra watershed
Use 'sudo apt autoremove' to remove them.
The following packages will be REMOVED:
  cuda* cuda-8-0* cuda-demo-suite-8-0* cuda-drivers* cuda-runtime-8-0* libcuda1-384*
  libcupti5.5* nvidia-384* nvidia-384-dev* nvidia-modprobe* nvidia-opencl-icd-384*
  nvidia-settings*
0 upgraded, 0 newly installed, 12 to remove and 56 not upgraded.
After this operation, 375 MB disk space will be freed.
Do you want to continue? [Y/n] y
(Reading database ... 219065 files and directories currently installed.)
Removing cuda (8.0.61-1) ...
Removing cuda-8-0 (8.0.61-1) ...
Removing cuda-demo-suite-8-0 (8.0.61-1) ...
Removing cuda-runtime-8-0 (8.0.61-1) ...
Removing cuda-drivers (384.111-1) ...
Removing libcupti5.5:amd64 (5.5.22-3ubuntu1) ...
Purging configuration files for libcupti5.5:amd64 (5.5.22-3ubuntu1) ...
Removing libcuda1-384 (384.111-0ubuntu1) ...
Purging configuration files for libcuda1-384 (384.111-0ubuntu1) ...
Removing nvidia-opencl-icd-384 (384.111-0ubuntu1) ...
Purging configuration files for nvidia-opencl-icd-384 (384.111-0ubuntu1) ...
Removing nvidia-384-dev (384.111-0ubuntu1) ...
Removing nvidia-384 (384.111-0ubuntu1) ...
Removing all DKMS Modules
Done.
update-alternatives: using /usr/lib/nvidia-384-prime/ld.so.conf to provide /etc/ld.so.con
f.d/x86_64-linux-gnu_GL.conf (x86_64-linux-gnu_gl_conf) in auto modeupdate-alternatives: using /usr/lib/nvidia-384-prime/ld.so.conf to provide /etc/ld.so.con
f.d/x86_64-linux-gnu_EGL.conf (x86_64-linux-gnu_egl_conf) in auto modeupdate-alternatives: using /usr/lib/nvidia-384-prime/alt_ld.so.conf to provide /etc/ld.so
.conf.d/i386-linux-gnu_GL.conf (i386-linux-gnu_gl_conf) in auto modeupdate-alternatives: using /usr/lib/nvidia-384-prime/alt_ld.so.conf to provide /etc/ld.so
.conf.d/i386-linux-gnu_EGL.conf (i386-linux-gnu_egl_conf) in auto modeupdate-alternatives: using /usr/lib/x86_64-linux-gnu/mesa/ld.so.conf to provide /etc/ld.s
o.conf.d/x86_64-linux-gnu_GL.conf (x86_64-linux-gnu_gl_conf) in auto modeupdate-alternatives: using /usr/lib/x86_64-linux-gnu/mesa-egl/ld.so.conf to provide /etc/
ld.so.conf.d/x86_64-linux-gnu_EGL.conf (x86_64-linux-gnu_egl_conf) in auto modeupdate-initramfs: deferring update (trigger activated)
Purging configuration files for nvidia-384 (384.111-0ubuntu1) ...
update-initramfs: deferring update (trigger activated)
Removing nvidia-modprobe (384.111-0ubuntu1) ...
Removing nvidia-settings (384.111-0ubuntu1) ...
Purging configuration files for nvidia-settings (384.111-0ubuntu1) ...
Processing triggers for libc-bin (2.23-0ubuntu10) ...
Processing triggers for man-db (2.7.5-1) ...
Processing triggers for initramfs-tools (0.122ubuntu8.11) ...
update-initramfs: Generating /boot/initrd.img-4.4.0-124-generic
W: Possible missing firmware /lib/firmware/ast_dp501_fw.bin for module ast
Processing triggers for desktop-file-utils (0.22-1ubuntu5.1) ...
Processing triggers for mime-support (3.59ubuntu1) ...
Processing triggers for gnome-menus (3.13.3-6ubuntu3.1) ...

13. 继续安装cuda

安装细节：

Installing the NVIDIA display driver...
Installing the CUDA Toolkit in /usr/local/cuda-9.1 ...
Installing the CUDA Samples in /home/linlf/tools/cuda9.1 ...
Copying samples to /home/linlf/tools/cuda9.1/NVIDIA_CUDA-9.1_Samples now...
Finished copying samples.

===========
= Summary =
===========

Driver:   Installed
Toolkit:  Installed in /usr/local/cuda-9.1
Samples:  Installed in /home/linlf/tools/cuda9.1

Please make sure that
 -   PATH includes /usr/local/cuda-9.1/bin
 -   LD_LIBRARY_PATH includes /usr/local/cuda-9.1/lib64, or, add /usr/local/cuda-9.1/lib64 to /etc/ld.so.conf and run ldconfig as root

To uninstall the CUDA Toolkit, run the uninstall script in /usr/local/cuda-9.1/bin
To uninstall the NVIDIA Driver, run nvidia-uninstall

Please see CUDA_Installation_Guide_Linux.pdf in /usr/local/cuda-9.1/doc/pdf for detailed information on setting up CUDA.

Logfile is /tmp/cuda_install_31995.log

14 最后，如果想更改cuda版本成9.1：

rm –rf /usr/local/cuda

ln -s /usr/local/cuda-9.1 /usr/local/cuda

参考： https://blog.csdn.net/u010821666/article/details/79957071