Profiles

This is a Core Concept page which offers an overview of key topics essential for understanding Profiles. Start here for foundational knowledge, then explore our guide and tutorials for more in-depth learning.

In this document you will learn what is profile and the differences between several profiles.
A profile defines the execution environment of your task, it consists of a pre configuration of Qarnot’s infrastructure. When you choose a profile, Qarnot automatically configures network access, system resource allocation (for example storage), licensing, hardware constraints like GPU acceleration.
A profile is generally associated with a software but you have the possibility to use custom docker image (see docker).

Fundamental principles

Typology

The naming convention for a profile is as follows:
software-[options] where software is the name of the software to use (ex: ansys-fluent, docker,…) and options one of:

Version: used for some software to specify the version to use
Network configuration (batch, local, network, ssh, cluster…): defines network rules, see Network for more information
Hardware configuration (gpu, nvidia,…): mostly to access GPU

Core Components

Each Profile sets hard configurations and some that you can override, here is a description of the main configurations and some generic parameters that you can override.

Disponibility

Public: the profile is available.
Payment required: the profile is available but you need to provide your payment information (generally profiles that require network access).
Private: generally profiles that are created specially for you (no other customer has access to it).
On request: this profile requires you to contact Qarnot to access to it.

Licensing

Free: no license is required to use the profile.
Included: the license fees are included in our pricing (generally for windows based software).
Bring your Own License: you are required to enter the information the task needs to access your license server (see license documentation).

Network Capabilities

Batch: Your task do not have access to any network, it is recommended to use it for high throughput computing where your task instances doesn’t need to communicate with each other.
Cluster or local: your tasks are able to interact with each other using local network or RoCE, it is recommended to use it for fluid simulations or data processing where several nodes need to interact altogether.
SSH: your task is accessible through ssh using the given IP and port.
Network: your task has access to internet.

If you need specific network capabilities (VPC, IP whitelisting…) please contact us.

Constants

The mandatory and optional parameters you need to use when configuring your profile. They can be:

Software specific parameters
- For more information on software specific constants please refer to the software pages.
Access parameters
- For example to use the right images from your docker registry, to configure the connexion to your license server or connect to a Database.
Configure the execution environment of your task
- DOCKER_PROGRES: metric used to customize progress bar.
- DOCKER_HYPERTHREAD: activate or deactivate CPU Hyper-thr.eading capabilities (advanced).
- DOCKER_USER: User to use when using your own docker image.

Software Specifics Profiles

Please check the Software section to learn more about software parameters and the blog to learn how to use them with the SDK.

Fore more information on how to use our platform, please take a look at our guides, tutorial, and documentation article:

Préférence de confidentialité