The following guest article from Intel explores how Intel Select Solutions is working to enable and simplify HPC for a broader set of users.
Simulation and modeling are the focus of one of the initial HPC workloads to be developed following the Intel Select Solutions protocols.
Building upon the success of Intel Cluster Ready and Intel Scalable System Framework (Intel SSF), Intel is facilitating the deployment of computing resources to non-traditional HPC users with the recently announced Intel Select Solutions for HPC.
Goals for all these programs focus on simplifying how users match their HPC applications to leading platforms and components, allowing more focus on productivity. Clusters based on Intel architecture and standard industry best practices could be designed and deployed without recourse to a do-it-yourself build for each cluster – a great savings in time, effort, and human resources.
These initiatives were not necessarily aimed at the massive data centers that belonged to the national labs or deep-pocketed universities. Both types of HPC data centers typically have cultivated infrastructure and management processes over many years. The new Intel Select Solutions build on the efforts of Intel Cluster Ready and Intel SSF to further enable how non-traditional HPC enterprises tap into the benefits of HPC without having to create and maintain specialized technical and administrative staff to support the cluster and its applications.
Intel Select Solutions allow Intel and its partners to develop and deploy the application-specific HPC solutions required to run today and tomorrow’s large and complex mix of workloads and applications.
Simulation and modeling are the focus of one of the initial HPC workloads to be developed following the Intel Select Solutions protocols. Intel and several OEMs and systems integrators are collaborating to facilitate deployment of common simulation and modeling applications such as computer-aided engineering (CAE), computational fluid dynamics (CFD) and finite element analysis (FEA)—applications that are particularly prevalent in the manufacturing industry. To support them at scale requires HPC systems with multicore processors, high-bandwidth fabrics, extensive storage and a full range of input/output capabilities. Assembling or upgrading to such a system can be a time-consuming, expensive project, entailing research, evaluation and deployment, as well as the integration and configuration of various components.
Intel is facilitating the deployment of computing resources to non-traditional HPC users with the recently announced Intel Select Solutions for HPC.
Intel Select Solutions for Simulation and Modeling simplify and speed up the purchase, configuration and deployment of a wide range of cluster configurations that ISVs and system integrators can tailor to their customers’ needs. A number of partners – ISVs and OEMs, with a specific interest in advanced simulation and modeling, are now in the final stages of developing applications for target industries. Rigorous benchmarking ensures that the system can support the latest modeling and simulation applications to deliver higher productivity and better performance.
Despite a flurry of initiatives like the National Network for Manufacturing Innovation and the US Advanced Manufacturing initiative, small and medium businesses and other non-traditional HPC users have been reluctant to adopt HPC and digital manufacturing for some very valid concerns: It’s too complicated, requires in-house expertise, and can add complexity to existing IT infrastructure that the business has relied on for years.
That may be changing. The Intel Select Solutions for Modeling and Simulation provide verified solutions with a pre-validated selection of components designed to meet the demands of HPC applications and workflows. These are now being created by Intel and its partners to provide organizations, and the ISVs, OEMs and consultants that support them, with affordable, easier-to-deploy solutions.
Learn more about Intel Select Solutions for Simulation and Modeling.
