AI Cloud
AI Cloud is Aalborg University's largest GPU cluster - perfect for machine learning and parallel processing tasks.
What is AI Cloud?
AI Cloud is a GPU cluster made up of a collecton of Nvidia and AMD GPU's, designed for processing GPU-demanding machine learning workloads.
The platform is accessed through a terminal application on the user's local machine. From here the user logs in to a front end node, where files management and job submission to the compute nodes takes place.
AI Cloud is best suited for batch processing workflows, ie. submitting jobs to a queueing mechanism and where they are processed on the compute nodes without user interaction. It is not designed for interactive development, where users occupy one or more GPU's while writing scripts.
Introduction to AI Cloud
Basic information and instructions for first-time users of AI Cloud.
Operating AI Cloud successfully involves being able to use the following:
- A Linux terminal environment.
- Slurm - a common mechanism for queueing HPC workloads.
- Singularity software containers.
- Training deep learning models for image classification and recognition tasks.
- Accelerating natural language processing with large-scale language models.
- Running high-resolution visual simulations with GPU-powered parallel computing.
- Performing video analysis for object detection and motion tracking in real-time.
Learn the Key Features
Discover the essential features of AI Cloud.
-
High-Performance GPU Cluster
AI Cloud provides access to a cluster of powerful NVIDIA GPUs, enabling the efficient processing of large datasets and complex models.
-
Containerization for Flexibility
Users operate within containerized environments, ensuring that software dependencies are consistently maintained across different computational nodes.
-
Workload Management
Jobs are managed via a queueing mechanism (Slurm), allowing for effective distribution across the available GPU resources.
-
Optimized for Parallel Processing
AI Cloud is specifically designed to handle large-scale, parallelized tasks, making it ideal for simulations, deep learning, and other high-throughput applications.
-
Comprehensive Documentation
Extensive resources are available to guide users through the setup and usage of AI Cloud, with a focus on accessibility for beginners.
-
Collaboration-Friendly Environment
AI Cloud supports file sharing and collaborative work among multiple users, facilitating teamwork on complex projects and enabling seamless integration of contributions from different researchers.
Important Information
Not for confidential or sensitive data
With AI Cloud you are only allowed to work with public or internal information according to AAU’s data classification model (classified as levels 0 and 1, respectively).
If you would like to work with confidential or sensitive data (classified as levels 2 and 3), then we support another HPC platform called UCloud.
Not suitable for CPU-only computational tasks
The powerful GPU processors allow users to process large datasets much more efficiently than would be the case with pure CPU processing - given that your application can be parallelised in a GPU compatible manner. At the same time, the AI Cloud platform is not designed for CPU-only computational tasks, and we have alternative recommended platforms, such as UCloud or Strato for those needs.
Review the terms and conditions
Before getting started, take a few moments to review the terms and conditions of using AI Cloud, and don't hesitate to reach out to our support team if you have any questions or concerns.