Smart Business Technology

Worried about the cost of a supercomputer?
A custom supercomputer doesn't have to crunch your budget. Inficron builds affordable enterprise HPC systems and offers business friendly cluster services and cluster hosting. Our scalable Single System Image (SSI) systems, based on OpenSSI, provide clients with high performance compute nodes that allow applications running on any one machine to access the resources of all the cluster nodes. High availability clusters leverage a fail-safe elected director to schedule jobs for compute nodes. Requests are load balanced over any number of designated cloud nodes that appear as a single IP system, making the supercomputer an ideal application server cluster or SQL cluster. High performance message passing is typically enabled using MVAPICH2 or OpenMPI depending on communication bandwidth requirements and links.

Inficron integrates OpenPBS, an open source portable batch scheduler, to provide job prioritization, queuing support, parallel jobs support, multiple scheduling algorithms, scripting and job submission. OpenPBS is extended with the Maui Scheduler to provide an array of scheduling policies, dynamic priorities, extensive reservations, and fairshare capabilities. MPIexec is also integrated to initialize parallel jobs from within PBS batch.

Clients are provided with comprehensive visibility into IT cluster statistics using Ganglia, with an integrated web front end, enabling administrators to view live or historical data such as CPU load averages or network utilization for all nodes. Building a supercomputer does not have to be difficult. Inficron does the work, so that clients can enjoy the performance with the option to host the high performance computing cluster with Inficron or have all hardware preconfigured and delivered to their premises, where Inficron engineers test for performance and stability with customer staff.

Learn more about how Inficron can help you scale a high performance computing cluster for your organization.