Qarnot gives you access to green parallel processing on fast high-end CPU. The computations are sent to Qarnot's infrastructure, heating people for free with Q.rad digital heaters.

With Qarnot's computing service, you can create or reuse the Docker images you already have to package and deploy your batch applications.

Use Qarnot's computing REST API to run your batch applications on-demand, in parallel, and at scale. You don't need to maintain complex infrastructure, Qarnot is doing the job for you.

Use cases

Qarnot's computing service has been designed to do batch processing and run bags of tasks. You can start in parallel a very a large number of similar tasks and retrieve your results faster than with any sequential approach.

Qarnot computing runs embarrassingly parallel applications or batches, that can be split in smaller tasks and ran at the same time on different nodes. As an example, Qarnot's platform is used for Monte-Carlo simulations in finance, 3D animation rendering, movie transcoding and more.

Parallel tasks

Using Qarnot computing

Running parallel tasks with Qarnot computing is done by using our REST API or one of our SDK (see Python SDK).

With the API, you create tasks, specify how many instances you want, upload the resources to process to buckets and download your results. Monitoring your tasks is also possible programmatically or through the Qarnot computing console.

What you need to start

To start to develop and run applications on Qarnot's computing infrastructure, you only need to create a Qarnot account.

API and SDKs

The easiest way to start is to use our Python SDK. If you are using another language, you can issue direct calls to our REST API.

APIs and SDKs Download Samples
Python SDK Github Hello world!, Working with files, Docker Hub, FFMpeg, Blender

Tasks and resources monitoring

In addition to the APIs and SDKs, Qarnot's console helps you monitor and follow the execution of your tasks. You can also delete unwanted tasks and manage your resources and results from there.

Qarnot console

The token that grants you access to Qarnot computing can be found by clicking on the top right buddy icon > API Token.

Qarnot computing workflow

The following schema shows a common development cycle, with an application hosted as a custom Docker image and few calls to our REST API.

Q.Ware workflow

  1. Prepare a Docker image that contains your application or find one on Docker Hub.

  2. Issue calls to our REST API to upload resource files and create a task with the desired number of instances.

  3. The Q.Ware fetches the requested Docker container for you.

  4. The Q.Ware finds suitable hardwares to run your container, including the Qarnot digital hearter (also known as Q.Rad).

  5. Once the compute has ended, your results are sent back to the Q.Ware. You can also be notified as soon as your task has ended.

  6. A new call to our REST API allows you to download your results.

Next steps