Download the full report.
This whitepaper is an insideHPC Special Research Report sponsored by Univa: Cloud Adoption for HPC: Trends and Opportunities.
Today many organizations solve their most difficult computational challenges by creating their own on-premises, compute-intensive, high performance data center. The latest hardware and software innovations have removed barriers that prevented many organizations from embracing HPC solutions.
Alternatively, cloud computing offers the flexibility of on-demand resources beyond the in-house data center. Integrating on-premises data centers with on-demand HPC resources in the cloud started out as a low-cost way to add flexibility. Now, as data centers struggle to handle expanding workloads involving simulation, AI, and deep learning, having the flexibility to provide dynamic capacity in support of innovation in these areas has become critical.
However, HPC in the cloud can become many times more expensive than installing and running highly efficient, low-latency, in-house compute-intensive clusters. But often, much of that cost is wasted due to over-provisioned or forgotten and not used resources, VMs left running, and other inefficiencies.
Managing the costs of running their HPC workflows in the cloud remains a major concern. With more enterprises extending their own on-premises clusters into the cloud with hybrid configurations, finding the right tools for managing and scheduling workflows and automatically controlling cost remains a critical challenge.
Download the new report courtesy of Univa, “Cloud Adoption for HPC: Trends and Opportunities” to explore the cloud adoption for HPC survey results.
