Many scientists and researchers need to perform CAE, CFD, artificial intelligence, deep learning, and a wide variety of high-performance computations and analyses that require the power of a supercomputer, but do not work in organizations that make HPC resources available. They would clearly benefit from access to high performance systems in order to speed their calculations, but even if they had access to HPC, likely lack the specialized knowledge required to configure it for their specific workflows. XTREME-D has invested years in tackling the problem of how to make HPC computing access easy, fast, efficient, and economical for all, over the cloud.
This technical deep dive will showcase the technology behind providing easy access to HPC and AI in the cloud from three perspectives: the hardware involved, the software necessary, and the automation required to allow for HPC resources to be configured on the fly. Cloud and hybrid computing are now desirable modes of computing for many, but continue to raise questions in regard to how to best implement these solutions, as well as how secure they are. This session will answer those questions.
We will describe XTREME-D’s innovative approach to helping users with limited HPC resources access and configure HPC-optimized hardware in a cloud environment, and discuss how a subscription-based usage model and the tools to support it provide unique flexibility and freedom of access to users. We will demonstrate this democratization of HPC through use cases from commercial enterprise, life sciences, and manufacturing.