Democratizing HPC with multicloud to accelerate engineering innovations

Today’s research is crucial because it fuels tomorrow’s innovations. Increasingly, the speed and magnitude of innovations rely on technology-powered research and engineering using high performance computing (HPC). That’s why democratizing HPC via the cloud—known as Cloud for HPC—can provide significant benefits to all of humankind. 

Cloud for HPC is helping to move HPC usage from fringe to mainstream, providing all researchers, engineers, and organizations access to this fundamental resource for innovation. There is, however, an important caveat: All Cloud for HPC solutions are not created equal. More on this in an upcoming section. First, let’s look at the organizational value of HPC-powered innovations.

The ROI of HPC-driven innovations

HPC provides value to most organizations, regardless of industry. A Hyperion study¹ estimates the return on investment (ROI) of HPC to be $44 of profit for every dollar of investment in HPC. “To put that into context, for an HPC system purchased at $100,000 by a private corporation, the analysis estimates that the profit will be around $4.4 million,” shares Hyperion.

Beyond its ROI, HPC is revolutionary for scientists and engineers. HPC is catalyzing breakthroughs for some of the world’s most complex and pressing challenges with incredible speed and precision. Here are a few outcomes from HPC’s groundbreaking capabilities:

• HPC modeling and simulation help engineers test jet engines without destroying a single $10-40 millionjet engine.²
• Previously, it took 10 years to fully sequence the DNA in a seawater sample. With HPC, it takes less than a week.³
• HPC is helping to hasten the green energy transition. For one company, Cloud for HPC reduced deployment time by 95% in designing and testing new engineering pumps, drill heads, and regulators.⁴

And this is just the beginning. Nearly every CEO today understands that leveraging artificial intelligence and machine learning (AI/ML) is a pathway to business viability. Lighthouse companies are leading the way while laggards are facing the risk of obsolescence. And changes are coming at warp speed. HPC usage will skyrocket as artificial intelligence/machine learning (AI/ML) and generative AI increase the need for more HPC resources.  

A shortage of HPC access

As computational demands and intensity rise, organizational IT challenges increase in kind. That’s because application portfolios are growing, and architectural requirements are becoming more diverse. Consequently, organizations are looking for more effective ways to deploy their infrastructure and leverage computing architectures on demand.

When it comes to HPC, traditional on-premises HPC often experiences very high utilization rates and is “maxed out” because it cannot scale elastically. In other words, organizations are approaching full use of their HPC clusters and there is not much spare capacity left. That means researchers and engineers must wait in a queue to do their work, delaying projects and stalling progress. For instance, longer-term workloads–such as COVID research in recent years–can take HPC priority, shutting out access for other projects. In other cases, project start dates can take top priority and worsen HPC access issues.

Another notable challenge creating a shortage of HPC access is affordability. Organizations that cannot afford on-premises HPC capabilities still need access to its computing power to solve pressing problems and stay competitive.

The much-needed answer to these two challenges can be found in the cloud.

HPC built for a multicloud world

Enter Cloud for HPC, which allows organizations to have the best of both worlds for HPC resourcing. Recently Dell Technologies partnered with Rescale to do just that and provide organizations with a hybrid on-premises, multicloud HPC solution.

The Dell HPC On Demand solution enables organizations to meet unpredictable spikes in demand alongside traditional on-premises HPC deployments. It does this with a blended hybrid cloud and on-prem approach, expanding HPC access and capacity as needed and on demand. For organizations that do not have the resources to deploy on-premises HPC, Cloud for HPC provides ready, affordable, and immediate access, democratizing HPC across industries, company sizes, and geographies.

However, not all Cloud for HPC solutions are created equal.

“Customers like the unique features of the HPC On Demand solution from Dell Technologies and Rescale. This includes features such as no vendor lock-in, a unified control plane to manage hybrid multi-cloud/on-premises resources, and the ability to burst seamlessly from on-premises to cloud,” shares Shervin Sammak, Ph.D., HPC Product Manager, Dell Technologies. “In short, HPC On Demand solution helps organizations effectively scale HPC capacity as their business grows.” 

Cloud for HPC: Things to know before you go

HPC On Demand provides distinct benefits to organizations with features such as:

User-centric. Gain anywhere, anytime access with no wait time using a simple, intuitive SaaS-like user interface.

Unlimited. Get any scale, any architecture, any application, any cloud, and any use case.

Connected. Access a distributed, globally connected network for multi-team, remote collaboration.

Intelligent. Optimize performance and costs across any workload type and automatically route workloads to the best architectures.

Automated. Achieve end-to-end visibility and policy-based control over teams, projects, and users with a single point for automation for multiple clouds, architectures, and regions.

“With HPC On Demand, we are helping organizations manage peak computing requirements, gain broad application support, and reduce IT maintenance burdens. This frees up resources to focus on other business priorities while providing on-demand access to HPC resources,” says Edward Hsu, Chief Product Officer,Rescale.

Accelerating the future with cloud for HPC

HPC On Demand is democratizing HPC and accelerating technology-powered innovations. The solution helps organizations with existing on-premises HPC and those without, by expanding access and providing scalability. The world’s next breakthrough research and engineering innovation may rely on the novel capabilities of Cloud for HPC and help shape a stronger future in areas like science, healthcare, aerospace, and the green energy transition. That’s a noble goal important to scientists, engineers, and every one of us.

***

Dell HPC On Demand provides customers with a flexible, cost-effective solution for their HPC needs. With the ability to seamlessly burst into the cloud and optimize spending across multicloud, customers can moderate their budget and avoid unplanned expenses. The absence of vendor lock-in also ensures customers have the freedom to choose the best resources for their needs.

Learn more about Dell HPC solutions at High Performance Computing | Dell USA

Rescale provides high performance computing built for the cloud. Since 2011, the revolutionary Rescale Platform has made it simple for scientists and engineers to harness the most advanced software and computing architectures for AI-driven research. Rescale helps organizations accelerate innovation in life sciences, transportation, energy, semiconductor, aerospace, manufacturing, and other industries. With Rescale HPC-as-a-service, organizations can access any application, on any architecture, and at any scale–all on the cloud of their choice. Learn more about the Rescale and Dell partnership.

^[1] https://www.hpcwire.com/2020/09/07/the-roi-on-hpc-44-in-profit-for-every-1-in-hpc/
^[2] https://www.delltechnologies.com/asset/en-us/products/ready-solutions/customer-stories-case-studies/dell-dxc-case-study.pdf
^[3] https://www.delltechnologies.com/asset/en-us/products/ready-solutions/customer-stories-case-studies/dell-oregon-state-case-study.pdf.external
^[4] https://rescale.com/solutions/by-industry/energy/