Hewlett Packard Enterprise (HPE) has unveiled a cutting-edge portfolio of high-performance computing (HPC) and artificial intelligence (AI) infrastructure solutions, including the Cray Supercomputing EX systems and two AI-optimized platforms tailored for large-scale language model training and multi-modal applications. These innovations are designed to accelerate advancements in scientific discovery and industry innovation globally.
“Organizations and nations focusing on sovereign AI projects are increasingly leveraging high-performance computing as a foundational resource for large-scale AI training, accelerating innovation and discovery,” said Trish Damkroger, Senior Vice President and General Manager of HPC & AI Infrastructure Solutions at the company. “Our expertise in delivering fully integrated, world-class HPC systems enables customers to deploy AI infrastructure quickly and efficiently, maximizing their return on investment.”
It’s latest HPC offerings, built on the its Cray Supercomputing EX platform, include groundbreaking advancements in compute, networking, storage, and software. Designed for research institutions and governments addressing complex global challenges, the portfolio features innovative fanless direct liquid cooling technology and state-of-the-art system architecture.
Key innovations include:
- HPE Cray Supercomputing EX4252 Gen 2 Compute Blade: Supports up to 98,304 cores per cabinet, offering high-density performance with 5th Gen AMD EPYC™ processors.
- HPE Cray Supercomputing EX154n Accelerator Blade: Houses up to 224 NVIDIA Blackwell GPUs, significantly reducing workload processing times.
- HPE Slingshot Interconnect 400: Offers 400 Gbps speeds, optimized for large workloads with low latency and congestion management.
- HPE Cray Supercomputing Storage Systems E2000: Doubles input/output performance, enhancing CPU and GPU utilization during operations.
- HPE Cray Supercomputing User Services Software: Streamlines system management, power efficiency, and workload execution.
The companyalso introduced the ProLiant Compute XD server series, optimized for AI training and deployment. The XD680 and XD685 models cater to enterprises and service providers building high-performance AI clusters, offering advanced cooling solutions, seamless integration, and enhanced security through HPE Integrated Lights-Out (iLO) technology.
As a global leader in supercomputing and AI, HPE continues to meet the evolving needs of customers by providing advanced systems that drive innovation in research, science, and enterprise applications. Their solutions will be showcased at SC24 in Atlanta, where the company highlights its leadership in next-generation computing technologies.
