Home

Qarnot

Qarnot provides unique low-carbon and sovereign cloud computing power.

Operating a distributed infrastructure hosted in France and Europe, the waste heat of our servers is locally recovered to heat buildings (offices, housings, public facilities, pools, warehouses...). Hence, Qarnot's expertise is twofold:

• highly efficient distributed computing,
• sustainable and digital heat.

You will find here all the resources you need to start using Qarnot's cloud computing platform.

Use cases

Qarnot gives you access to green computing power on high-end CPUs and GPUs. Using the cloud platform, you can deploy your computations on-demand, in parallel and at scale. This is typically suited for batch-computing: you can start a very large number of computations at the same time and retrieve their results faster than with any sequential approach.

Step by step deployment

For a quick introduction to Qarnot's features, we advise you to jump to the Getting Started section of this documentation. It will help you through your first steps on Qarnot's cloud platform.

For a broader glimpse, follow along this guide and learn how to start a computation on Qarnot's cloud.

1. Create an account

Create a Qarnot account  and retrieve your personal authentication token.

Find how to manage your Qarnot Account on Tasq to access your authentication token and gain insights on your usage of the platform.

2. Choose your software

Qarnot uses Docker  containerization to easily package and deploy applications in the cloud. To run computations on the platform, you need to select a Docker image containing your application and its dependencies. You can either:

• use a Docker image optimized by Qarnot: explore our integrated software
• find an off-the-shelf Docker image: a wide variety of Docker images already exist and are freely available on public repositories,
• use your own custom Docker image.

To learn how to use one or the other solution presented above look for examples on our blog.

If you want to run Qarnot-integrated software, you do not need to be familiar with Docker. However, to run other software or combination of tools, we advise you to consult Docker documentation.

3. Launch a computation

Choose the method that best suits your needs to upload your data to the platform and launch your computations as computing tasks:

• develop your pipeline with a SDK  (Python, C# or Node.js) or the CLI,
• send direct requests to the API,
• launch a task in your web browser through our graphical interface.

Each of these methods is detailed in the Launching a computation from scratch section.

4. Manage your data

On Qarnot's cloud platform, data is stored as objects in buckets using S3 (Simple Storage Service) protocol. As for launching computations (see above), there are several ways to manage your data:

• through one of the SDKs (Python, C# or Node.js) or the CLI,
• using our graphical interface,
• with any S3 compatible third party software.

Basic usage of the SDKs, CLI and GUIs to manage data is presented in the corresponding pages of section Launching a computation from scratch. To find advanced usages and examples, refer to the section Managing your data.

The above are the easiest ways to manage your files and folders on Qarnot's cloud. However, for large amounts of data, you might need to use a third party software.

5. Monitor your computations

At any given time, you can monitor your current and past tasks through one of the following tools:

• a SDK (Python, C# or Node.js) or the CLI,
• one of our two graphical interfaces (Tasq or Render),
• the REST API.

Carbon metrics

Qarnot provides you with Carbon Facts to get reliable insights on the carbon footprint of your computations. Learn more about carbon metrics.

Security

Learn about Qarnot security policy.