Skip to main content

Using the GPU in the cloud

Here is an example workflow for how to use the GPU in the cloud to render your videos using EC2 instances.

Update Remotion​

Ensure you are using Remotion v4.0.248 or later for this guide to work.

Launch an EC2 instance​

1
 Ensure you are allowed to use a GPU

You might need to ask AWS for a limit increase for the number of GPUs you can use.
You can do this in the AWS console.
Go to "Service Quotas" -> "AWS Services" -> "Amazon Elastic Compute Cloud (Amazon EC2)" -> "Running On-Demand G and VT instances" -> "Request increase at account-level".
You may also click here to go to the page directly for the us-east-1 region.

2
 Launch an EC2 instance

Click here to launch an EC2 instance on us-east-1.

Select "Browse more AMIs", search for ami-053b0d53c279acc90, select the "Community AMIs" tab and select the image with the right AMI (ubuntu/images/hvm-ssd/ubuntu-jammy-22.04-amd64-server-20230516).

note

It might take up to 30 seconds until the search finds the AMI and it is not the first result.

We recommend the g4dn.xlarge size - note that this instance costs $375 per month with the default configuration.
If you get a message "Subscribing to AMI is taking longer than expected", it is normal. You may need to wait a few minutes.

3
 Configure instance:

Once you connected to the instance, run the following commands:

Upgrade the Linux Kernel to v6
bash
sudo bash -c "apt update && export DEBIAN_FRONTEND=noninteractive && export NEEDRESTART_MODE=a && apt upgrade -y && reboot"

The instance will restart and therefore disconnect. Wait a few moments and then reconnect.

Install the Linux Dependencies for Remotion:

Install Linux Dependencies
bash
sudo apt install -y \
  libnss3 \
  libdbus-1-3 \
  libatk1.0-0 \
  libasound2 \
  libxrandr2 \
  libxkbcommon-dev \
  libxfixes3 \
  libxcomposite1 \
  libxdamage1 \
  libcups2 \
  libgbm-dev \
  libpangocairo-1.0-0 \
  libatk-bridge2.0-0

Set up GPU drivers:

Install libvulkan
bash
sudo apt install -y build-essential libvulkan1
Install GPU drivers
bash
DRIVER_URL="https://us.download.nvidia.com/tesla/535.104.12/NVIDIA-Linux-x86_64-535.104.12.run"
DRIVER_NAME="NVIDIA-Linux-driver.run"
wget -O "$DRIVER_NAME" "$DRIVER_URL"
sudo sh "$DRIVER_NAME" --disable-nouveau --silent
rm "$DRIVER_NAME"
Configure startup service
bash
echo '[Unit]
Description=Run nvidia-smi at system startup
[Service]
ExecStart=/usr/bin/nvidia-smi
Type=oneshot
RemainAfterExit=yes
[Install]
WantedBy=multi-user.target' | sudo tee /etc/systemd/system/nvidia-smi.service
sudo systemctl enable nvidia-smi.service
sudo systemctl start nvidia-smi.service
Install Node.js
bash
sudo mkdir -p /etc/apt/keyrings
curl -fsSL https://deb.nodesource.com/gpgkey/nodesource-repo.gpg.key | sudo gpg --dearmor -o /etc/apt/keyrings/nodesource.gpg
NODE_MAJOR=20
echo "deb [signed-by=/etc/apt/keyrings/nodesource.gpg] https://deb.nodesource.com/node_$NODE_MAJOR.x nodistro main" | sudo tee /etc/apt/sources.list.d/nodesource.list
sudo apt-get update
sudo apt-get install nodejs -y

Render a video with the GPU

Clone a Remotion GPU demo
bash
git clone https://github.com/remotion-dev/gpu-scene
cd gpu-scene
npm i
npx remotion gpu --chrome-mode="chrome-for-testing" --gl=vulkan # Verify content is hardware accelerated
npx remotion render --chrome-mode="chrome-for-testing" --gl=vulkan # Render a test video

With --chrome-mode="chrome-for-testing", a flavor of Chrome is used that emulates a display, which can make use of the GPU.
With --gl=vulkan, Chrome uses the right OpenGL renderer.

note

Warnings such as vkCreateInstance() failed: -7 and Failed to create and initialize Vulkan implementation. may show up, however, the render should be significantly faster.

See also​