If you get empty result, it can be concluded the cuDNN library is not present on the system.ĥ. We run this command the get the currently installed cuDNN library $ locate cudnn | grep "libcudnn.so." | tail -n1 | sed -r 's/^.*\.so\.//'Īlternative command: $ find /usr -name "*libcudnn.so.*" | head -n1 | sed -r 's/^.*\.so\.//' Check if there is existing cuDNN library installed Use this command to install the package: $ sudo apt-get install mlocateĪlternatively, we can substitute locate command with find command as follows: $ find /usr -name "*libcupti.so.*" | head -n1 | sed -r 's/^.*\.so\.//'Ĥ. Note: if you see the error message “ bash: locate: command not found”, you may need to first install mlocate package. We run this command to get the installed CUDA profiling tools version: $ locate cupti | grep "libcupti.so." | tail -n1 | sed -r 's/^.*\.so\.//'Įmpty result means that the CUDA profiling tools library has not been installed. Check if there is existing CUDA profiling tools installed If you see the error message “ bash: nvcc: command not found”, it implies that CUDA has not been installed or previous installation was not successful.ģ. We run this command to get the graphics driver version: $ nvidia-smi | grep "Driver Version" | awk '' | cut -c2. Check the NVIDIA graphics driver that is currently installed. Tip: The commands shown in this section are taken from the command cheatsheet written in this article.ġ. Pre-Installation Checksīefore we proceed with the installation, we need to check and confirm the current state of installed software components by running a series of commands as outlined below. This post will cover three scenarios of CUDA 9.2 installation: 1) fresh installation, 2) install to upgrade by removing old version, 3) install to upgrade and keep multiple versions. Newer version will may come soon or later and here we are left with the bogging question: “How can we upgrade safely without clobbering the currently working system?” Moreover, we may also wonder if there is a mechanism to rollback the change and live with current setup while recognizing that it’s not yet the time to upgrade. It is also safe to assume that CUDA 9.2 will not be final version. As with other software that evolves, NVIDIA released CUDA 9.2 back in May. In the previous post, we’ve proceeded with CUDA 9.1 installation on Ubuntu 16.04 LTS.
0 Comments
Leave a Reply. |
Details
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |