Hewlett Packard Enterprise (HPE) will deliver a comprehensive high-performance computing (HPC) and Artificial Intelligence (AI) portfolio for the exascale era.
The addition of Cray, Inc., which HPE recently acquired, bolsters HPE’s HPC and AI solutions to now encompass an end-to-end supercomputing architecture across compute, interconnect, software, storage and services, delivered on premises, hybrid or as-a-Service.
HPE’s latest HPC and AI end-to-end portfolio is comprised of the following services, software, compute, interconnect and storage capabilities:
Making HPC and AI accessible as-a-Service
Cray and Microsoft have collaborated to allow access to HPC technology and systems for Cray in Microsoft Azure offerings. HPE plans to expand customer choice for customers by providing its HPC and AI solutions through HPE GreenLake, an as-a-Service offering from HPE.
Purpose-built software to manage all types of data
- HPC and AI systems run like a cloud – Cray System Management Software, built on a container-based, multi-tenant architecture, enables converged HPC and AI workloads to run simultaneously on a single system, providing administrators and developers with a modern cloud experience.
- Simplifying and optimizing workloads – HPE is simplifying application development and management across complex HPC and AI workloads that consist of diverse processors and accelerators with the Cray Programming Environment (CPE). CPE provides a fully supported developer environment and integrated software suite, offering compilers and programming languages, tools and libraries, that maximizes programmer productivity, application scalability and performance. It is designed to allow easy porting of existing applications with minimal recoding, while minimizing changes to existing programming models and simplifying developer transition to new hardware paradigm.
- Extended support for HPE Performance Cluster Management: The HPE Performance Cluster Management (HPCM) software provides complete provisioning, management, and monitoring for clusters scaling to 100,000 nodes. The software enables fast system setup from bare-metal, hardware monitoring and management, image management, software updates and power management. As part of the new portfolio, future planned releases of HPCM will expand system support to include new Cray systems.
Compute optimized for any environment, any size
HPE is adding Cray Supercomputers to its portfolio. Starting at a single server and expanding to the largest supercomputers, the combined portfolio can address the supercomputing needs of any data center. The new Cray systems will be based on the Shasta architecture, which will be used to deliver the first three U.S. exascale systems.
High performance interconnect for growing data traffic
HPE is offering Cray Slingshot, a redefined HPC interconnect, to address demands for higher speed and lower latency to process the explosion of new data and convergence of HPC and AI workloads. Slingshot starts with 200 gbps throughput and adds unique congestion control capabilities to deliver the performance and scale required to drive numerous data-intensive workflows simultaneously on a single system.
Faster, intelligent storage for modern HPC
High performance storage: The recent introduction of Cray ClusterStor E1000, a storage solution and a new type of solution for HPE’s HPC portfolio, supports ongoing data growth and converged workloads. ClusterStor meets storage requirements while significantly reducing the need for additional storage drives through its intelligent new data management software.
Efficient high touch service for tailored experiences
HPE delivers a personalized experience with HPE Pointnext services, a consulting and technical services offering that provides collaborative efforts to advise, design and accelerate deployment of any digital experience or roadmap to meet an organization’s goals. HPE is further bolstering these offerings with Cray’s services, which include support features such as installation, troubleshooting, system maintenance, system administration, and system optimization.
HPE will showcase several solutions at the annual Supercomputing Conference 2019 (SC 19) in Denver, Colorado, such as the new HPE ProLiant DL385 and DL325 Gen10 Plus servers, based on 2nd Gen AMD EPYC processors which enhance performance and efficiency for vital workloads like HPC, and an AI Ops solution as part of an R&D collaboration with the National Renewable Energy Laboratory (NREL) to improve energy efficiency, sustainability and resiliency in supercomputing data centers.