First, install desktop environment and vncsever
sudo apt-get install ubuntu-desktop gnome-panel gnome-settings-daemon metacity nautilus gnome-terminal
sudo apt-get update
sudo apt-get upgrade
sudo apt-get install vnc4server
Use vncserver to start and vncserver -kill :1 to kill vncserver session.
Optionally use vncserver -geometry 1920x1080 to specify resolution
Update content of ~/.vnc/xstartup, need to restart vncserver to take effect
# Uncomment the following two lines for normal desktop:
# unset SESSION_MANAGER
# exec /etc/X11/xinit/xinitrc
[ -x /etc/vnc/xstartup ] && exec /etc/vnc/xstartup
[ -r $HOME/.Xresources ] && xrdb $HOME/.Xresources
xsetroot -solid grey
vncconfig -iconic &
x-terminal-emulator -geometry 80x24+10+10 -ls -title "$VNCDESKTOP Desktop" &
x-window-manager &
gnome-panel &
gnome-settings-daemon &
metacity &
nautilus &
Forward port, using the following gcloud command to forward port 5901 (to local port 5911) as well as to login
gcloud compute --project "some-proj-123456" ssh --zone "us-east1-c" "instance-2" --ssh-flag "-L 5911:127.0.0.1:5901"
Use VNC viewer (https://www.realvnc.com/) to open 5911 port
Follow instruction here: https://cloud.google.com/compute/docs/gpus/add-gpus
Driver installation script for Ubuntu 16.04
#!/bin/bash
echo "Checking for CUDA and installing."
# Check for CUDA and try to install.
if ! dpkg-query -W cuda; then
# The 16.04 installer works with 16.10.
curl -O http://developer.download.nvidia.com/compute/cuda/repos/ubuntu1604/x86_64/cuda-repo-ubuntu1604_8.0.61-1_amd64.deb
dpkg -i ./cuda-repo-ubuntu1604_8.0.61-1_amd64.deb
apt-get update
apt-get install cuda -y
fi
Verification
$ nvidia-smi
Wed Jul 12 23:20:10 2017
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 375.66 Driver Version: 375.66 |
|-------------------------------+----------------------+----------------------+
| GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
|===============================+======================+======================|
| 0 Tesla K80 Off | 0000:00:04.0 Off | 0 |
| N/A 38C P0 58W / 149W | 0MiB / 11439MiB | 100% Default |
+-------------------------------+----------------------+----------------------+
+-----------------------------------------------------------------------------+
| Processes: GPU Memory |
| GPU PID Type Process name Usage |
|=============================================================================|
| No running processes found |
+-----------------------------------------------------------------------------+
Installing cuDNN:
Now getting cuDNN here https://developer.nvidia.com/cudnn (need to log in, and download cuDNN 5.1 not 6.0)
Issue the following commands to install it (https://askubuntu.com/questions/767269/how-can-i-install-cudnn-on-ubuntu-16-04)
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*
First install python 3.6 version anaconda release
Then
pip install tensorflow-gpu keras tqdm