Cray Converged System Architecture and Software for Machine Learning and Data Analytics at Extreme Scale
Abstract
Extreme Scale AI presents a number of infrastructure challenges, both in terms of hardware components (networks, processors and storage) and the varied resource-intensiveness of different stages in the workflow (data preparation vs. model training). This presentation will introduce Cray's vision for a converged system architecture and explore the software environment that enables workflows associated with data analytics and machine learning at extreme scale.
Case studies from work done at Cray and at scale on Cray systems will be presented.- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2018
- Bibcode:
- 2018AGUFMIN11A..09C
- Keywords:
-
- 1906 Computational models;
- algorithms;
- INFORMATICSDE: 1914 Data mining;
- INFORMATICSDE: 1942 Machine learning;
- INFORMATICSDE: 1978 Software re-use;
- INFORMATICS