Scalable Data Processing and Management for Converged High-performance and Intelligent Computing

Scalable Data Processing and Management for Converged High-performance and Intelligent Computing

Monday, May 13, 2024 3:00 PM to 3:20 PM · 20 min. (Europe/Berlin)
Hall H, Booth L01
HPC Solutions Forum
Compiler and Tools for Parallel ProgrammingParallel File SystemsResource Management and SchedulingScalable Application Frameworks

Information

Over the past decades, the large availability of data has been increasingly pushed forward from a multitude of sources, including numeric simulations, observations and instrumentations. The massive data volume profoundly revolutionizes scientific discoveries and engineering advances. And, rapid growth of intelligent computing has been further transforming the data-driven computing landscape, and rapidly converges with high performance computing to shape new paradigms. Big data and large-scale computing unprecedently necessitate novel processing and management techniques. Towards the essential demands, we develop converged ecosystem platform to empower the capabilities of various domains such as healthcare and material science. Particularly, to support scalable data management and analytics, we craft efficient file system and framework infrastructures on TH-2 supercomputer. Regarding the evolving landscape, we are actively exploring full-stack solutions for emerging architectures like supercomputer internet to enrich the potentials and impacts of converged computing.
HPC Solutions Forum Questions
Are technologies and configurations for AI and HPC converging or diverging? Is it possible to serve both adequately and efficiently in the same environment?What can be done to make HPC and AI more efficient?
Format
On-site