High Performance Computing Gets Its Own HyperHub

Nimbix, a pioneer in the cloud high performance computing (HPC) space, today launched HyperHub, a self-service HPC application marketplace that enables supercomputing across multiple cloud environments.

With HyperHub, engineers and scientists can select from an ecosystem of prebuilt applications and workflows and run them on any device, any cloud or on-premises infrastructure. Nimbix, an eight-year old, Dallas-based technology company, is among the early cloud HPC players and is known as a cloud platform and infrastructure provider that’s purpose-built for HPC workloads.

Nimbix’s Steve Hebert

“The HyperHub marketplace is extremely unique for the HPC market in the cloud landscape,” Nimbix CEO Steve Hebert told Channel Futures. “Nimbix has taken the work it’s done over the years, curating a large catalog of high-performance applications that the Nimbix cloud has offered as a point-and-click capability to its customers, and extended it beyond the Nimbix cloud to any infrastructure — whether on-premises to a compute cluster managed by the customer or to the cloud provider of their choice or preferred geography.”

HyperHub, with its well-oiled catalog of applications and workflows, solves a couple of unique challenges. One is the challenge that many companies have regarding the complexity of the use cases and the software applications that are used to do their computing — whether that’s domains such as simulation or data science — that coupled with having to install them on complex infrastructure, either on-premises or in the cloud.

“So, this [HyperHub] is a really important from an IT simplification process,” said Hebert.

A second technology aspect that’s important, according to Hebert, is that Nimbix pioneered the use of container technology in HPC. “That technology area [containers] is extremely hot right now across the IT landscape and HyperHub enables that wave of containerized applications as part of the platform,” he said.

HyperHub provides users instant access to commercial HPC software, accelerated workflows and open-source applications built by Nimbix, its partners and independent software vendors (ISVs). With point-and-click simplicity, users can launch any of these optimized apps and workflows to solve a variety of compute-intensive problems, without worrying about installing, configuring and tuning their own complex software stacks from scratch. HyperHub is powered by Nimbix’s JARVICE XE platform.

Users who go to the HyperHub marketplace will find simulation tools such as Ansys Fluent commonly used in the automotive, aerospace, manufacturing and energy space, where this is staple software for their use cases, for example. “That catalog and suite of simulation use cases will be in HyperHub,” said Hebert.

Another example of use cases that customers will find in HyperHub are in the ML, deep learning and AI space, such as Tensor open source OpenFlow, MXNet, and various flavors of these ML and cognitive types of software.

Users will also find more traditional HPC and supercomputing codes for various use cases such as finite element analysis to physics types of computational applications.

In 2018, Nimbix introduced JARVICE XE, extending the capabilities of its JARVICE operating system beyond the Nimbix Cloud to any cloud infrastructure, whether single, hybrid…

…or multicloud, including field-programmable gate array (FPGA) and graphics processing unit (GPU) accelerated clusters.

The HPC market focuses on use cases such as science and engineering that is becoming more pervasive across a larger set of customers. It includes everything down from the application level to the cloud infrastructure.

“When we started cloud HPC was very new and with the advent of deep learning and AI, and many of these very computationally-intensive use cases, it’s become more pervasive as an IT challenge for many companies,” said Hebert.

Nimbix is a small cloud player, and the only HPC-only focused player, in a universe made up of competitors such AWS, Google Cloud, and Microsoft.