HPE Unveils Next-Generation Cray Supercomputing Portfolio to Accelerate AI Advancements

The newly unveiled portfolio includes three multi-partner, multi-workload compute blades, enhanced unified management software, and robust high-performance interconnect support, positioning HPE as a leader in supercomputing for research institutions, government entities, and large enterprises. Trish Damkroger, Senior Vice President and General Manager of HPC and AI Infrastructure Solutions at HPE, emphasized the platform's role in meeting the needs of global organizations reliant on supercomputing for improved performance across diverse workloads.

The portfolio expansion follows the recent introduction of the HPE Cray Supercomputing GX5000, designed specifically for breakthroughs in the converged AI and HPC era. This advanced platform is complemented by the HPE Cray Supercomputing Storage Systems K3000, recognized as the industry’s first factory-built storage system with integrated Distributed Asynchronous Object Storage (DAOS) open-source software, enhancing performance capabilities.

The HPE Cray Supercomputing GX5000 has already garnered attention from prestigious institutions, including the High-Performance Computing Center of the University of Stuttgart (HLRS) and the Leibniz Supercomputing Centre (LRZ) of the Bavarian Academy of Sciences and Humanities. Both institutions have selected the GX5000 for their upcoming supercomputers, the Herder and Blue Lion, respectively, signifying strong industry demand for HPE’s cutting-edge technology.

Prof. Dr.-Ing. Michael Resch, director at HLRS, noted the profound impact of HPE's supercomputing systems on scientific discovery and technology design, while Prof. Dr. Dieter Kranzlmüller, chairman of LRZ, praised the GX5000's energy-efficient design, emphasizing its potential to transform research workflows by integrating AI methods alongside traditional modeling.

The new multi-partner, multi-workload processing blades offer unparalleled performance density, supporting flagship GPUs such as NVIDIA’s Rubin platform and AMD’s Instinct™ MI430X, as well as next-gen AMD EPYC™ processors. The three processing blades include:

HPE’s new Supercomputing Management Software introduces capabilities for multi-tenant, virtualized, and containerized environments, ensuring flexible deployment of AI and HPC resources. The software enhances monitoring of power usage and energy efficiency, critical for managing operational costs in high-demand settings.

The introduction of HPE Slingshot 400 for GX5000-based systems further enhances interconnect capabilities, providing reduced latency and improved bandwidth for high-performance applications.

The HPE Cray Supercomputing Storage Systems K3000 leverages the HPE ProLiant Compute DL360 Gen12 server, delivering high-speed data transfer rates essential for demanding AI applications. The availability of various drive sizes and configurations ensures scalability tailored to project requirements.

HPE emphasizes its commitment to customer support, offering services that include application performance optimization, global implementation, and continuous operational assistance. As a recognized leader in the supercomputing domain, HPE aims to empower organizations to achieve their most ambitious goals.

Industry leaders have echoed the significance of HPE’s advancements. Travis Karr from AMD highlighted the collaborative effort in developing high-density compute technologies, while Dion Harris from NVIDIA underscored the role of HPE's systems in accelerating scientific discovery. Dr. Earl Joseph from Hyperion Research emphasized the growth potential in HPC and AI sectors, suggesting that HPE's innovations will lead to enhanced research capabilities and societal advancements.

In summary, HPE's latest Cray supercomputing portfolio represents a strategic move to address the growing computational needs of AI and HPC, ensuring organizations can leverage next-gen technology for groundbreaking scientific and industrial outcomes.