Building on our progress in the exascale era, HPE is on a mission to deliver our leadership technologies to accelerate all customers’ innovation and discovery. At ISC 2023, HPE invites you to visit with our experts onsite to learn how you can leverage the latest supercomputing systems for faster time-to-results to achieve your boldest scientific and business goals. Key solutions highlighted in our booth (#G705) will include:

Life sciences – See how HPE GreenLake for HPC helps to optimize the resource usage of life sciences workloads in a cloud infrastructure while satisfying all the privacy, security, compliance, and regulations of this industry.

Manufacturing – What HPC and AI convergence looks like – Integrating existing user workflows like Crash Simulation aaS with no or limited changes in the customer’s day-to-day practices, provide the user-experience of HPC cloud services with an HPE GreenLake for HPC model.

Developing Generative AI models on converged HPC + AI clusters – See how the HPE Machine Learning Development Environment enables both HPC and AI workloads to dynamically and efficiently share the same accelerated compute platforms to run large-scale AI workloads, such as training GPT-like models alongside HPC workloads.

Quantum computing – Learn about technology advancements that can empower phenomenal breakthroughs with simulations and integrations to achieve quantum advantage today.

Last year, at ISC 2022, the 59th edition of the TOP500^[1] revealed the HPE Cray EX supercomputer, Frontier, to be the first true exascale machine with an HPL score of 1.102 Exaflops. This massive machine, which accounted for 25% of the total flops on the Top500^[2] when it was introduced, required the floorspace of over two basketball courts to house the required compute and storage systems. Since then, HPE has delivered a new generation of HPE Cray supercomputers including the HPE Cray EX2500, the HPE Cray XD2000, and the HPE Cray XD6500, all of which bring the technology of HPE Cray supercomputing in configurations and sizes enterprises can and have adopted.

The Top500^[3] list unveiled at Supercomputing 2023 last November showed some interesting trends from an HPE perspective. With 101 systems on the Top500, HPE led the field in total performance delivered with 45% of the cumulative flops on the list. What was more revealing was the configurations of the systems over time. Of the 80 HPE systems on the November 2022 TOP500 introduced before 2022, 20 of them contained accelerators. Of the 21 HPE systems on the Top500 introduced in 2022, 19 of them contained accelerators signaling a strong shift to AI and mixed HPC & AI workloads. It is also interesting to note that this trend is not pervasive across the Top500 with the percentage of accelerated systems actually decreasing for non HPE systems. HPE alone accounted for over one-half of the accelerated systems in the Top500 introduced last year highlighting a strong preference for HPE technologies for AI and mixed HPC & AI workloads.

At HPE, we find that the market is choosing our solutions for their large-scaling AI implementations. Hewlett Packard Enterprise is reimagining what infrastructure can do by bringing together leading compute, storage, and software development technologies that are foundational for the use and optimization of AI-at-scale.

Our goal is simple and powerful:

• Reduce complexity and cost of model development and training at scale
• Eliminate enormous operational difficulties of deploying and managing an AI infrastructure
• Mitigate any other common last mile challenges of operationalizing AI

HPE platforms such as the first member of the HPE Cray XD6500 family, the HPE Cray XD670, are optimized for enterprises looking to deploy AI-at-scale. The breakthrough server infrastructure for the enterprise provides high agility and throughput across all system components to enable high performance, flexible expansion, and AI acceleration. With a highly dense architecture, the system delivers the performance needed to run the most demanding applications.

The HPE Cray EX2500 delivers a complete integrated supercomputing cluster in a size most datacenters can accommodate. Supporting the same compute and network blades as its larger predecessor, the HPE Cray EX4000, it can also deliver GPU accelerated nodes designed for AI-at-scale workloads.

The Cray ClusterStor E1000 Storage System is purpose-engineered to meet the demanding input/output requirements of supercomputers and HPC clusters in a very efficient way. The E1000 parallel storage solution typically achieves the given HPC storage performance requirements, significantly reducing the number of storage drives. That means HPC users with a fixed budget for the HPC system can spend more of their budget on CPU/GPU compute nodes, accelerating time-to-insight.

While obviously important, performance alone does not make a great, large-scaling AI solution. The HPE Machine Learning Development Environment enables companies to build transformative AI applications at scale by removing the complexity and cost associated with ML model development, making it easier for IT administrators to set up, manage, secure, and share AI compute clusters. With HPE Machine Learning Development Environment, ML practitioners can train models faster using state-of-the-art distributed training without changing their model code, automatically find high-quality models, get more from their GPUs with smart scheduling as well as reduce cloud GPU costs by seamlessly using spot instances, and track and reproduce their work with experiment tracking.

An ideal large scaling AI solution includes not only performant hardware and development environment, but also the storage, networking, and cluster management required for straightforward operation. HPE Machine Learning Development System is a complete solution that allows enterprises to focus on innovation, not infrastructure. Validated and pre-configured hardware, software, and services for model development and training make it faster to deliver compute and technical resources to keep up with the demands of AI/ML. Enterprises can perform ML across accelerated computing clusters without rewriting code or updating infrastructure, enabling teams to iterate and collaborate on inventive models and launch the best ones into production.

Finally, and importantly, at HPE, we have transformed all aspects of supercomputing to enhance computing power without generating the prohibitive increases in energy that can potentially lead to higher operational costs and lasting environmental damage.

With all our systems, we continue to rethink the way we incorporate efficient power and cooling techniques, energy-aware software, machine learning AI, and more to optimize energy efficiency. We offer high-performance computing as-a-Service via HPE GreenLake for HPC, which provides access to dedicated, powerful computing and analytics capabilities while minimizing upfront costs, reducing overprovisioning, and enabling customers with the elastic capacity to scale quickly. We also demonstrate sustainable technology solutions leadership by working closely with customers to keep products in use longer by refurbishing and remarketing equipment.

Supercomputing plays a key role in our lives with potential to realize a more sustainable future for everyone. At HPE, we continue to push the boundaries of what is possible by maximizing the performance of the world’s most powerful supercomputers while keeping energy efficiency central to our design, products, services, and operations. We are proud to have 101 of our systems on the November 2023 Green500[4] list of the world’s most energy efficient supercomputers, with 7 HPE supercomputers comprising the Top 10 of the list.

HPE is a platinum sponsor of International Supercomputing 2023 in Hamburg, Germany, from May 22-May 24. Join us to see how we are pushing the boundaries of what is possible with HPE Cray supercomputing technologies and AI. We look forward to showing you the full breadth of what you can do with our integrated solutions.

Planning to attend? Register today for a one-on-one meeting meeting with HPE supercomputing and AI experts. Visit us HPE at booth #G705.

---

[1] June 2022 Top500 list https://www.top500.org/lists/top500/2022/06/
[2] June 2022 Top500 list https://www.top500.org/lists/top500/2022/06/, Frontier had 1.102 exaflops, total for Top500 was 4.403 exaflops
[3] November 2023 Top500 list https://www.top500.org/lists/top500/2022/11/
[4] November 2022 Green500 list https://www.top500.org/lists/green500/2022/11/