Optimize your AI power with scheduling

AiQu is a versatile tool that will help to maximize your organizations AI-initiatives.

We help our clients get there faster

AiQu brings the future with less of a hazzle

The advent of Deep Learning and the explosion of GPU based systems for “AI” development has created new challenges for both Data Scientists and IT Operations alike. The tools traditionally used for scheduling, and orchestrating, CPU-based HPC-environments are no longer the optimal choice for GPU-based systems. Organizations want their highly valuable investments in both hardware and personnel to be used in the most efficient way possible.

Therefore, CGit has developed AiQu a versatile tool that will help to maximize your organizations AI-initiatives.

Challenges

Simplicity and ease of use is key

When we started development of AiQu we wanted a tool that got rid of the complexity that many of the, still popular, schedulers are weighed down by. We know from experience working with some of the most advanced AI teams in Scandinavia, that simplicity and ease of use is key. We also wanted the footprint of our scheduler to be as small as possible to be able to install it in environments that might be constrained, like on switches for instance.

We know that many tools are not NVLink, or NVswitch, aware. If we could add that to our scheduler it would be a big win for efficiency and control. We also wanted to add control over already queued or running jobs as well as handle prioritizations, project management and access control.

AiQu will solve these challenges for you.

It is up to you

Ease of use and Control

AiQu will give you unprecedented visibility and control over your workloads and queues. Projects or single users can be given “tokens” to be used as collateral to run jobs on certain GPU’s, number of GPU’s or prioritization even.

Choose between a user-friendly GUI or do all your work via CLI, it is up to you. AiQu uses the same syntax as you do locally on your client.

Get started

AiQu is free to try

Try in a limited version and can either be purchased as a perpetual license to be installed locally, or as a SaaS service from CGit.

Key features

Autonomous load

Keeping your valuable compute-resources busy is one of AiQu’s prime tasks. Autonomous Load makes this possible by loading scheduled jobs without human interference while also making sure that you are in complete control.

Smart Resource Management

AiQu will optimize your resource management. No more struggling with oversubscription of resources. AiQu monitors the resources used and will recommend the most cost-effective number of GPUs for your workload.

Integrations

API-first design so that AiQu can be integrated into any workflow thereby optimizing your projects (and your valuable time).

Security

AiQu keeps users separated from compute resources, and data, thereby reducing the risk of any malicious intent and keeping your projects safe and compliant.

Calendar view and Project Workflow*

AiQu provides a calendar tool and Project Workflow that will give you and your teams excellent visibility and the ability to plan your workflow in an optimal way. *Coming H1 2021

Technical Support from experienced engineers

The team behind AiQu has more experience in AI-infrastructure and Deep Learning workloads, than almost anyone else on the market, and will provide market leading support when needed.

Platform Agnostic

AiQu can schedule jobs to any hardware platform anywhere.

Small Footprint

AiQu is designed from the ground up to be as efficient as possible and can run on very limited resources.

Integrate with your tool of choice

AiQu integrates with all tools that can be containerized.
Or why not build your own model from scratch?