WIXLIB

Chapter 3.PACKAGE MANAGER INSTALLATION3.1. OverviewThe Package Manager installation interfaces with your system's package managementsystem. When using RPM or Deb, the downloaded package is a repository package.Such a package only informs the package manager where to find the actual installationpackages, but will not install them.If those packages are available in an online repository, they will be automaticallydownloaded in a later step. Otherwise, the repository package also installs a localrepository containing the installation packages on the system. Whether the repository isavailable online or installed locally, the installation procedure is identical and made ofseveral steps.Distribution-specific instructions detail how to install SUSEUbuntuL4TFinally, some helpful package manager capabilities are detailed.These instructions are for native development only. For cross development, see theCross-build environment for ARM section.3.2. Redhat/CentOS1. Perform the pre-installation actions.2. Satisfy DKMS dependencywww.nvidia.comNVIDIA CUDA Getting Started Guide for LinuxDU-05347-001 v6.5 8

Package Manager InstallationThe NVIDIA driver RPM packages depend on other external packages, such asDKMS and libvdpau. Those packages are only available on third-party repositories,such as EPEL. Any such third-party repositories must be added to the packagemanager repository database before installing the NVIDIA driver RPM packages, ormissing dependencies will prevent the installation from proceeding.3. Address custom xorg.conf, if applicableThe driver relies on an automatically generated xorg.conf file at /etc/X11/xorg.conf.If a custom-built xorg.conf file is present, this functionality will be disabled and thedriver may not work. You can try removing the existing xorg.conf file, or adding thecontents of /etc/X11/xorg.conf.d/00-nvidia.conf to the xorg.conf file. The xorg.conffile will most likely need manual tweaking for systems with a non-trivial GPUconfiguration.4. Install repository meta-data sudo rpm --install cuda-repo- distro - version . architecture .rpm5. Clean Yum repository cache sudo yum clean expire-cache6. Install CUDA sudo yum install cudaIf the i686 libvdpau package dependency fails to install, try using the followingsteps to fix the issue: yumdownloader libvdpau.i686 sudo rpm -U --oldpackage libvdpau*.rpm7. Add libcuda.so symbolic link, if necessaryThe libcuda.so library is installed in the /usr/lib{,64}/nvidia directory. For preexisting projects which use libcuda.so, it may be useful to add a symbolic link fromlibcuda.so in the /usr/lib{,64} directory.8. Perform the post-installation actions.3.3. Fedora1. Perform the pre-installation actions.2. Address custom xorg.conf, if applicableThe driver relies on an automatically generated xorg.conf file at /etc/X11/xorg.conf.If a custom-built xorg.conf file is present, this functionality will be disabled and thedriver may not work. You can try removing the existing xorg.conf file, or adding thecontents of /etc/X11/xorg.conf.d/00-nvidia.conf to the xorg.conf file. The xorg.conffile will most likely need manual tweaking for systems with a non-trivial GPUconfiguration.3. Satisfy Akmods dependencyThe NVIDIA driver RPM packages depend on the Akmods framework which isprovided by the RPMFusion free repository. The RPMFusion free repository mustbe added to the package manager repository database before installing the NVIDIAwww.nvidia.comNVIDIA CUDA Getting Started Guide for LinuxDU-05347-001 v6.5 9

Package Manager Installationdriver RPM packages, or missing dependencies will prevent the installation fromproceeding.4. Install repository meta-data sudo rpm --install cuda-repo- distro - version . architecture .rpm5. Clean Yum repository cache sudo yum clean expire-cache6. Install CUDA sudo yum install cudaThe CUDA driver installation may fail if the RPMFusion non-free repository isenabled. In this case, CUDA installations should temporarily disable the RPMFusionnon-free repository: sudo yum --disablerepo "rpmfusion-nonfree*" install cudaIf also installing the gpu-deployment-kit package, the cuda and gpudeployment-kit packages should be either installed using separate instances ofyum: sudo yum install cuda sudo yum install gpu-deployment-kitOr, installed while also specifying the cuda-drivers package: sudo yum install cuda cuda-drivers gpu-deployment-kitIf a system has installed both packages with the same instance of yum, some drivercomponents may be missing. Such an installation can be corrected by running: sudo yum install cuda-driversIf the i686 libvdpau package dependency fails to install, try using the followingsteps to fix the issue: yumdownloader libvdpau.i686 sudo rpm -U --oldpackage libvdpau*.rpm7. Add libcuda.so symbolic link, if necessaryThe libcuda.so library is installed in the /usr/lib{,64}/nvidia directory. For preexisting projects which use libcuda.so, it may be useful to add a symbolic link fromlibcuda.so in the /usr/lib{,64} directory.8. Perform the post-installation actions.3.4. SLES1. Perform the pre-installation actions.2. Install repository meta-data sudo rpm --install cuda-repo- distro - version . architecture .rpm3. Refresh Zypper repository cachewww.nvidia.comNVIDIA CUDA Getting Started Guide for LinuxDU-05347-001 v6.5 10

Package Manager Installation sudo zypper refresh4. Install CUDA sudo zypper install cudaThe driver is provided in multiple packages, nvidia-gfxG03-kmp-desktop, nvidiagfxG03-kmp-default, nvidia-gfxG03-kmp-trace, and their Unified Memory variants.When installing cuda, the correct driver packages should also be specified. Withoutdoing this, zypper will select packages that may not work on the system. Run thefollowing to detect the flavor of kernel and install cuda with the appropriate driverpackages: uname -r3.4.6-2.10- flavor sudo zypper install cuda nvidia-gfxG03-kmp- flavor \nvidia-uvm-gfxG03-kmp- flavor 5. Add the user to the video group sudo usermod -a -G video username 6. Install CUDA Samples GL dependenciesThe CUDA Samples package on SLES does not include dependencies on GL and X11libraries as these are provided in the SLES SDK. These packages must be installedseparately, depending on which samples you want to use.7. Perform the post-installation actions.3.5. OpenSUSE1. Perform the pre-installation actions.2. Install repository meta-data sudo rpm --install cuda-repo- distro - version . architecture .rpm3. Refresh Zypper repository cache sudo zypper refresh4. Install CUDA sudo zypper install cudaThe driver is provided in multiple packages, nvidia-gfxG03-kmp-desktop, nvidiagfxG03-kmp-default, nvidia-gfxG03-kmp-trace, and their Unified Memory variants.When installing cuda, the correct driver packages should also be specified. Withoutdoing this, zypper will select packages that may not work on the system. Run thefollowing to detect the flavor of kernel and install cuda with the appropriate driverpackages: uname -r3.4.6-2.10- flavor sudo zypper install cuda nvidia-gfxG03-kmp- flavor \nvidia-uvm-gfxG03-kmp- flavor 5. Add the user to the video groupwww.nvidia.comNVIDIA CUDA Getting Started Guide for LinuxDU-05347-001 v6.5 11

Package Manager Installation sudo usermod -a -G video username 6. Perform the post-installation actions.3.6. Ubuntu1. Perform the pre-installation actions.2. Install repository meta-dataWhen using a proxy server with aptitude, ensure that wget is set up to use thesame proxy settings before installing the cuda-repo package. sudo dpkg -i cuda-repo- distro version architecture .deb3. Update the Apt repository cache sudo apt-get update4. Install CUDA sudo apt-get install cuda5. Perform the post-installation actions.3.7. L4T1. Perform the pre-installation actions.2. Install repository meta-data sudo dpkg -i cuda-repo- distro version architecture .deb3. Update the Apt repository cache sudo apt-get update4. Install CUDA Toolkit sudo apt-get install cuda-toolkit-6-55. Add the user to the video group sudo usermod -a -G video username 6. Perform the post-installation actions.3.8. Additional Package Manager CapabilitiesBelow are some additional capabilities of the package manager that users can takeadvantage of.www.nvidia.comNVIDIA CUDA Getting Started Guide for LinuxDU-05347-001 v6.5 12

Package Manager Installation3.8.1. Available PackagesThe recommended installation package is the cuda package. This package will install thefull set of other CUDA packages required for native development and should cover mostscenarios.The cuda package installs all the available packages for native developments. Thatincludes the compiler, the debugger, the profiler, the math libraries,. For x86 64patforms, this also include NSight Eclipse Edition and the visual profiler It also includesthe NVIDIA driver package.On supported platforms, the cuda-cross-armhf package installs all the packagesrequired for cross-platform development on ARMv7. The libraries and header files ofthe ARMv7 display driver package are also installed to enable the cross compilationof ARMv7 applications. The cuda-cross-armhf package does not install the nativedisplay driver.The packages installed by the packages above can also be installed individually byspecifying their names explicitly. The list of available packages be can obtained with: yum --disablerepo "*" --enablerepo "cuda*" list available zypper packages -r cuda cat /var/lib/apt/lists/*cuda*Packages grep "Package:"# RedHat & Fedora# OpenSUSE & SLES# Ubuntu3.8.2. Package UpgradesThe cuda package points to the latest stable release of the CUDA Toolkit. When a newversion is available, use the following commands to upgrade the toolkit and driver: sudo yum install cuda sudo zypper install cuda sudo apt-get install cuda# RedHat & Fedora# OpenSUSE & SLES# UbuntuThe cuda-cross-armhf package can also be upgraded in the same manner.The cuda-drivers package points to the latest driver release available in the CUDArepository. When a new version is available, use the following commands to upgrade thedriver: sudo yum install cuda-drivers sudo zypper install cuda-drivers \nvidia-gfxG03-kmp- flavor \nvidia-uvm-gfxG03-kmp- flavor sudo apt-get install cuda-drivers# RedHat & Fedora# OpenSUSE & SLES# UbuntuSome desktop environments, such as GNOME or KDE, will display an notification alertwhen new packages are available.To avoid any automatic upgrade, and lock down the toolkit installation to the X.Yrelease, install the cuda-X-Y or cuda-cross-armhf-X-Y package.Side-by-side installations are supported. For instance, to install both the X.Y CUDAToolkit and the X.Y 1 CUDA Toolkit, install the cuda-X.Y and cuda-X.Y 1 packages.www.nvidia.comNVIDIA CUDA Getting Started Guide for LinuxDU-05347-001 v6.5 13

Chapter 4.RUNFILE INSTALLATIONThis section describes the installation and configuration of CUDA when using thestandalone installer.4.1. Pre-installation SetupBefore the stand-alone installation can be run, perform the pre-installation actions.4.2. PrerequisitesIf you have already installed a standalone CUDA driver and desire to keep using it,you need to make sure it meets the minimum version requirement for the toolkit. Thisrequirement can be found in the CUDA Toolkit release notes. With many distributions,the driver version number can be found in the graphical interface menus underApplications System Tools NVIDIA X Server Settings. On thecommand line, the driver version number can be found by running /usr/bin/nvidiasettings.4.3. ContentsThe standalone installer can install any combination of the NVIDIA Driver (that includesthe CUDA Driver), the CUDA Toolkit, or the CUDA Samples. If needed, each individualinstaller can be extracted by using the -extract /absolute/path/to/extract/location/. The extraction path must be an absolute path.The CUDA Toolkit installation includes a read-only copy of the CUDA Samples. Theread-only copy is used to create a writable copy of the CUDA Samples at some otherlocation at any point in time. To create this writable copy, use the cuda-installsamples-6.5.sh script provided with the toolkit. It is equivalent to installing theCUDA Samples with the standalone installer.www.nvidia.comNVIDIA CUDA Getting Started Guide for LinuxDU-05347-001 v6.5 14

Runfile Installation4.4. Graphical Interface ShutdownExit the GUI if you are in a GUI environment by pressing Ctrl-Alt-Backspace. Somedistributions require you to press this sequence twice in a row; others have disabled italtogether in favor of a command such as sudo service lightdm stop. Still othersrequire changing the system runlevel using a command such as /sbin/init 3 Consultyour distribution's documentation to find out how to properly exit the GUI. This step isonly required in the event that you want to install the NVIDIA Display Driver includedin the standalone installer.4.5. InstallationTo install any combination of the driver, toolkit, and the samples, simply executethe .run script. The installation of the driver requires the script to be run with rootprivileges. Depending on the target location, the toolkit and samples installations mayalso require root privileges.By default, the toolkit and the samples will install under /usr/local/cuda-6.5 and (HOME)/NVIDIA CUDA-6.5 Samples, respectively. In addition, a symbolic linkis created from /usr/local/cuda to /usr/local/cuda-6.5. The symbolic link iscreated in order for existing projects to automatically make use of the newly installedCUDA Toolkit.If the target system includes both an integrated GPU (iGPU) and a discrete GPU(dGPU), the --no-opengl-libs option must be used. Otherwise, the openGL library usedby the graphics driver of the iGPU will be overwritten and the GUI will not work. Inaddition, the xorg.conf update at the end of the installation must be decli

NVIDIA CUDA Getting Started Guide for Linux DU-05347-001_v6.5 1 Chapter 1. INTRODUCTION CUDA is a parallel computing platform and programming model invented by NVIDIA. It enables dramatic increases in computing performance by harnessing the power of the graphics processing unit (GP