Render on AWS

Create a p2.xlarge instance, which I found is the best balance between performance and price. Make sure it’s ubuntu based.

Once ready, SSH into it and prepare it to render scenes:

sudo apt-get update
 
# need the same gcc-compiler as the kernel uses, or so NVidia says
sudo apt install -y gcc=4:9.3.0-1ubuntu2 make libgl1-mesa-glx libxi6 libxrender1
 
# download and install the NVidia drivers
wget https://us.download.nvidia.com/XFree86/Linux-x86_64/470.94/NVIDIA-Linux-x86_64-470.94.run
chmod +x NVIDIA-Linux-x86_64-470.94.run
sudo ./NVIDIA-Linux-x86_64-470.94.run
 
# manually go through the install options
 
# install latest Blender
sudo snap install blender --classic
 
# create the enable-gpu.py script to force the use of GPU
cat << 'EOF' > enable-gpu.py
import bpy
 
prop = bpy.context.preferences.addons['cycles'].preferences
prop.get_devices()
prop.compute_device_type = 'CUDA'
 
for device in prop.devices:
    if device.type == 'CUDA':
        device.use = True
 
bpy.context.scene.cycles.device = 'GPU'
 
for scene in bpy.data.scenes:
    scene.cycles.device = 'GPU'
EOF

To upload files into it: Uploading files to AWS instances.

To render a scene:

# start a detached screen so you won't lose progress when disconnecting
screen -S myRenderingSession
 
# start blender rendering
blender -b blenderFile.blend -P enable-gpu.py -a
 
# if you had lost the connection and need to reconnect to that screen session
screen -d -r
 
# if you need to start blender again but don't want to start from scratch, but instead, frame 10
blender -b blenderFile.blend -P enable-gpu.py -s 10 -a