Workshops Archive

AI on HPC: Performance Engineering, Challenges and Opportunities


Workshop: AI on HPC: Performance Engineering, Challenges and Opportunities

Authors: Siavash Ghiasvand (TUD Dresden University of Technology), Vijeta Sharma (Sweden AI Factory), and Ajeet Ram Pathak (Norwegian University of Science and Technology)

Abstract: How can AI workloads be engineered for optimal performance in modern HPC environments? As AI models grow in scale and complexity, high performance computing (HPC) systems are becoming essential for training and deploying these workloads. However, traditional HPC architectures, designed for batch processing, present unique challenges for AI/ML applications — such as dynamic workloads, high I/O demands, and parallelization intricacies. This workshop explores the intersection of AI and HPC, focusing on performance optimization, scalability, and portability. Topics include AI workload characterization, hybrid AI/HPC workflows, efficient execution strategies, and infrastructure co-design. By fostering collaboration between researchers, practitioners, and system developers, this workshop aims to advance best practices and innovations in AI-enabled scientific computing, aligning with SC26’s themes of exascale readiness and performance portability.

Website: https://ai-on-hpc.github.io/sc26


Back to AI on HPC: Performance Engineering, Challenges and Opportunities Archive Listing

Back to Full Workshop Archive Listing