Published November 30, 2024 | Version v1

D2.4 Integrated AI-Scheduler Prototype

  • 1. University of Göttingen
  • 2. Cardiff University School of Chemistry

Contributors

Description

The Integrated AI-Scheduler Prototype is designed to dynamically optimize and manage job schedul-

ing across diverse compute environments, including HPC, cloud, and edge systems. The AI-Scheduler

employs advanced algorithms to select the most efficient scheduling strategy based on real-time job

and node characteristics.

This deliverable outlines the prototype’s implementation, demonstrating its capabilities through

simulations and real-world use cases. The report covers the methodologies applied, the tools and

models developed, engineering process to validate and optimize the system output and the key

achievements made in this phase. The findings showcase the effectiveness of the AI-Scheduler in

improving energy efficiency, resource utilization, and job throughput across different environments.

Files

D2_4IntegratedAI_SchedulerPrototype.pdf

Files (2.0 MB)

Name Size Download all
md5:00fbeb4bf43ee4638b6aa36ba5ab0737
2.0 MB Preview Download

Additional details

Funding

European Commission
DECICE - Device-Edge-Cloud Intelligent Collaboration framEwork 101092582