HPC just got (even) easier

[Max_Alt]

Real-time data analysis is now within everyone’s reach. Max Alt explains how HPE GreenLake for HPC is making it possible.

What do cutting-edge prototypes, sophisticated financial models, and new cures for diseases have in common?

I’ll give you a clue: they all depend on the same technology.

The kind that swallows vast amounts of data and churns out operational efficiencies before breakfast—with number-crunching far beyond any human and, to be honest, most machines.

Most machines, but not all.

You guessed it. I'm talking about high performance computing (HPC).

From saving money to saving lives, HPC is revolutionizing almost everything we care about. Together with AI and machine learning, HPC is now virtually essential across various industries—from manufacturing to oil and gas, financial services to life sciences and healthcare.

Why? If data holds the answers, everything depends on how fast you can unlock them. Hence the need for real-time analytics is driving demand for HPC at an unprecedented scale. More computing power means deeper insights, better decisions, and faster time to value. It's that simple.

But hold on.

Isn't HPC infrastructure expensive and complex?

Yes. The skills to operate, tune, and manage HPC in-house are hard to come by—and harder to afford. Meanwhile, many organizations are overpaying in search of the same performance from the public cloud.

“We just landed a new contract. HPC could really improve our computer-aided-design—but we can’t risk that much capital.”

“We need HPC to accelerate our drug discovery, but not all the time. We don’t want to get locked into overprovisioning in the cloud.”

Sound familiar?

Whatever business you're in, it's the same problem. That's why HPE decided to change the game.

HPE is democratizing HPC 

HPE GreenLake for HPC is an on-premises, fully managed, end-to-end solution that makes it easier and faster for customers to deploy HPC and AI workloads. And we've just made it even more accessible.

• Lower barrier to entry. High performance doesn't have to mean high costs. HPE GreenLake for HPC swaps hefty upfront investment for a flexible pay-per-use model. You can now start with just 10 nodes—allowing you to test and scale with zero risk. Extra capacity costs the same as regular capacity, so you can meet spikes in demand without any surprises. There's no data movement, no egress fees, and no expensive app refactoring to worry about. Essentially, you get as much power as you need when you need it. Predictable, manageable, affordable.
• Infrastructure plus expertise. HPE GreenLake for HPC is a fully managed service: you get the latest infrastructure, and we look after it for you. Our experts quickly tailor the service to your needs and manage everything behind the sciences. The service is available on-premises or in a colocation facility near your data, whatever works best. What do you do? Put all that power to good use.

What's new? 

Here’s a quick recap for anyone that missed the latest announcements. HPE has added new, purpose-built HPC capabilities to help organizations quickly tackle the most demanding compute and data-intensive workloads, power AI and ML initiatives, and speed time to insight.

As I mentioned above, you can now start with just ten nodes, making it easy to dip into HPC—and see where it takes you…

• Expanded GPU capabilities. Integrating with the new HPE Apollo 6500 Gen10 system, you can accelerate compute and advance data-intensive projects using the NVIDIA A100, A40, and A30 Tensor Core GPUs in increments of 2-4-8 accelerators.
• HPE Slingshot. Leveraging Cray networking technology, this is the world’s only high-performance Ethernet fabric designed for HPC and AI solutions. HPE Slingshot enables organizations to manage data-intensive and AI workloads with higher speed and congestion control.
• HPE Parallel File System Storage. A scalable, high-performance storage solution delivering unprecedented throughput for broader HPC and AI needs.
• Multi-cloud connector APIs. Offering the ability to programmatically orchestrate HPC workflows on a diverse pool of computing resources such as other HPE GreenLake for HPC or public clouds. You get more elasticity, scalability, and tools to optimize usage of disaggregated resources. You can also improve collaboration by connecting to other projects between multiple sites, removing silos.

Fact sheet: HPE GreenLake for High Performance Computing

HPC as a service in action

What does this look like in practice? HPE recently helped a European engineering design manufacturer to transform their HPC capabilities. Previously, they were struggling to manage and scale a sprawling global network of HPC clusters. With multiple engineers spread across different geographies and business units, the business was suffering from inconsistent workflow and sub-optimal network performance—with a growing "Shadow IT" presence to meet user demand and circumvent Central IT Ops. With HPE GreenLake for HPC, they now enjoy an optimized and consistent user experience and improved overall performance, underpinned by a scalable, future-proof platform that drives operational and cost efficiencies.

Join us at ISC

It's an exciting time for HPC, and I can't wait to see what incredible innovations are around the corner. We're currently gearing up for ISC High Performance 2022 in Hamburg from May 29–June 2 2022. It's the first live ISC post-pandemic, so it's going to be a big one!

Come along, and I'll be delighted to give you a live demo of HPE GreenLake for HPC.

Max Alt
Hewlett Packard Enterprise

twitter.com/hpe_hpc
linkedin.com/showcase/hpe-ai/
hpe.com/info/hpc