High Performance Computing on the Cloud.
We live in a highly competitive and dynamic world where the time to bring products to market can be a core differentiator. Enterprises need to build, test, tweak and retest models for their products depending on the business they are in, whether these are related to infusing artificial intelligence (AI) in their business models, designing a new car and understanding the impact of airflow or performing counterparty credit risk modeling. These models need enormous computing power to produce timely and accurate results.
In the past, large corporations and governments with the financial muscle to use supercomputers could tap into the massive computing power for solving problems around meteorology, life sciences, financial services and fluid dynamics.
High Performance Computing on the Cloud — also known as HPC on Cloud — is democratizing the near supercomputing grade power on cloud by making it an operating expense play for organizations and creating a level playing field for all, including startups.
The first requirement of HPC on Cloud is enterprise-grade compute, where IBM has leading-edge capabilities. The offering leverages IBM Cloud VPC featuring Intel Xeon processors. Another important component is a scheduler — such as Spectrum LSF and Spectrum Symphony — that can orchestrate jobs by utilizing resources in the most effective and efficient manner.
IBM Symphony is now available on IBM Cloud
A few months ago, we released IBM Spectrum LSF on IBM Cloud. Today, we announce the release of IBM Symphony on IBM Cloud. It will help you set up an end-to-end HPC system with an option of one or more schedulers — like IBM Symphony on Cloud — with a few simple clicks.
The IBM Spectrum Symphony® software delivers powerful, enterprise-class management for running compute-intensive and data-intensive distributed applications on a scalable, shared grid. It accelerates dozens of parallel applications for faster results and better utilization of all available resources. With IBM Spectrum Symphony, you can improve IT performance, reduce infrastructure costs and expenses, and quickly meet business demands.
Why Cloud for HPC?
- Hybrid and agile: Enterprises often struggle with availability of resources on their on-premises HPC setup because there are always critical jobs competing for resources, especially if there is a seasonal peak or any other trigger (such as a natural disaster). Cloud gives you an option to burst your workloads with full agility.
- Operating expense vs. capital expense: Cloud, in general, solves the IT Capex problem by converting it into an Opex expense. It is even more relevant in thecase of High Performance Computing since the upfront capital expense for an on-prem HPC setup is high.
- Time to market: Because resources are available on demand on the cloud, enterprises can rapidly develop their business models without any infrastructure constraints. This expedites the overall delivery plan.
- Democratize resources: Today, HPC is not limited to the few organizations that have large budgets. With a small budget outlay, a startup having a great idea can start performing complex calculations that require massive compute resources. This helps them develop new ideas and products quickly.
- Access to a large ecosystem: IBM has a very rich suite of applications readily available on the IBM Cloud. You can easily and rapidly use these applications to perform various activities on the HPC resources.