diff --git a/dependency_hints.txt b/dependency_hints.txt new file mode 100644 index 0000000000000000000000000000000000000000..90dc2d2f31f67eae4d4aab39a34618529cea2026 --- /dev/null +++ b/dependency_hints.txt @@ -0,0 +1,107 @@ +The following notes are the commands I used to install the NVIDIA bianry drivers, CUDA, cuDNN, and ROS Indigo. +After doing these commands, openpose and its dependencies compiled for me quite easily. + +sudo apt-get update +sudo apt-get upgrade + +# NVIDIA Titan XP Binary Driver Installation + +Installed the following driver for NVIDIA Titan XP: http://www.geforce.com/drivers/results/117741 +Following the instructions at https://ubuntuforums.org/showthread.php?t=2081649 using the binary instructions + +1) Uninstall previous nvidia drivers +sudo apt-get purge nvidia-current nvidia-settings +2) sudo service lightdm stop + +3) ctrl-alt-f1 +4) Login +5) Then switch to the directory where your drivers are + +cd Downloads + +6) Then install the driver: + +sudo ./N +# Fix no-TTY terminal on ctl-alt-F1 with Nvidia drivers + +https://forums.linuxmint.com/viewtopic.php?t=168108 + +sudo nano /etc/default/grub + +Uncomment the lines + +#GRUB_TERMINAL=console + +#GRUB_GFXMODE=640x480 + +And add nomodeset voa: + +GRUB_CMDLINE_LINUX_DEFAULT="quiet splash nomodeset" +Cuda 8.0 Install + +Download the .run installer from https://developer.nvidia.com/cuda-downloads (NOT THE DEB file) + +Perform the pre-installation instructions http://docs.nvidia.com/cuda/cuda-installation-guide-linux/index.html#pre-installation-actions + +(so far, I am able to skip the pre-installation instructions) + +Following : http://www.insiderattack.net/2014/12/installing-cuda-toolkit-65-in-ubuntu.html + +# Install CUDA via: + +sudo chmod a+x cuda_8.0....run +sudo ./cuda_8.0.....run + +Say no to installing the driver (since we already did it). Say yes to everything else. + +Then perform the post installation instructions: (http://docs.nvidia.com/cuda/cuda-installation-guide-linux/index.html#post-installation-actions) + +Add these to your ~/.bashrc + +export PATH=/usr/local/cuda-8.0/bin${PATH:+:${PATH}} + +export LD_LIBRARY_PATH=/usr/local/cuda-8.0/lib64\ + ${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}} + + +See if you can compile the examples: + +cd ~/NVIDIA_"examples" + +make -j8 + +I had an error with g++ so I just installed it via: + +sudo apt-get install g++ + + +# Install cudann + +https://askubuntu.com/questions/767269/how-can-i-install-cudnn-on-ubuntu-16-04 + +step 0: Install cuda from the standard repositories. (See How can I install CUDA on Ubuntu 16.04?) + +Step 1: Register an nvidia developer account and download cudnn here (about 80 MB) + +Step 2: Check where your cuda installation is. For the installation from the repository it is /usr/lib/... and /usr/include. Otherwise, it will be /usr/local/cuda/ or /usr/local/cuda-<version>. You can check it with which nvcc or ldconfig -p | grep cuda + +Step 3: Copy the files: + +$ cd folder/extracted/contents +$ sudo cp -P include/cudnn.h /usr/include +$ sudo cp -P lib64/libcudnn* /usr/lib/x86_64-linux-gnu/ +$ sudo chmod a+r /usr/lib/x86_64-linux-gnu/libcudnn* + +For a local install of cuda (not using deb) + + 82 cd Downloads/cuda/ + + 84 sudo cp -P include/cudnn.h /usr/local/include/ + + 90 sudo cp -P lib64/libcudnn* /usr/local/cuda-8.0/lib64/ + + 91 sudo chmod a+r /usr/local/cuda-8.0/lib64/libcudnn* + +# Install ROS Indigo + +http://wiki.ros.org/indigo/Installation/Ubuntu \ No newline at end of file