In the ever-evolving landscape of HPC and AI, the intersection of cutting-edge technology and groundbreaking innovation is where possibilities become reality.
Advances in addressing complex challenges such as climate modeling, aviation innovation, drug discovery, nuclear defense simulations and cancer and disease diagnosis are now possible thanks to todayโs powerful supercomputers, fast interconnect and HPC/AI software.
As the trusted global supercomputing leader[i], HPE continues to remain at the forefront of innovation with a newly expanded portfolio. Today, we introduce systems that use the latest generation of CPUs and GPUs, bringing the most advanced supercomputing technologies to organizations with the right scale to enable advancements in some of the most exciting areas of discovery.
HPE introduced several exciting announcements this morning at Supercomputing 23 in Denver, Colorado. For those that canโt be there in person, hereโs a highlight of the buzz on the show floor:
Introducing the HPE Cray Supercomputing XD665 System
Last year, HPE introduced a new product line that makes supercomputers more accessible for enterprises by introducing systems at a lower price point in a form factor small enough to fit into the average data center. Since then, HPE has expanded that portfolio and continues to introduce more systems that use the latest generation processors from AMD and NVIDIA.
Industries such as health and life sciences, financial services, manufacturing and public sector require superior modeling, high-end simulation capabilities and the ability to manage accelerated computing applications. The HPE Cray Supercomputing XD665 offers these organizations leading-edge and sustainable HPC and AI capabilities on a system running up to four NVIDIA H100 Tensor Core GPUs and direct liquid cooling (DLC).
The HPE Cray Supercomputing XD665 system is an ideal solution for heavily parallel and sequential workloads, making it a balanced solution for both simulation and training needs. This system is available now with flexible deployment options including HPE GreenLake.
Machine learning model training made easier
Another exciting announcement includes significant updates to HPE Machine Learning Development System, a comprehensive solution that helps mitigate IT complexity with a dedicated, high-performance AI infrastructure for Generative AI model development. Designed to reduce the complexity of AI/ML infrastructure deployment and management, the HPE Machine Learning Development System leverages scalable configurations for improved efficiency of operations and GPU density for the most challenging AI workloads.
The HPE Machine Learning Development System helps customers to easily and securely integrate Generative AI capabilities with a purpose-built solution that is preconfigured, fully installed, and performant out-of-the-box, and is comprised of a concise combination of HPE Cray hardware and the HPE Machine Learning Development Environment software.
New blades unlock new business value and discovery
The ability to compute massive datasets quickly and efficiently is critical to business outcomes. Complex simulations and AI training workloads, including natural language processing (NLP) and computer vision, can take weeks to process on traditional blades. Running AI workloads on an exascale-class supercomputer can reduce that time down to days, or even hours.
HPE announces two new accelerator blade offerings based on the HPE Cray Supercomputing EX4000 and EX2500. These new blades offer the power of a supercomputer in a footprint small enough to be housed in an enterprise data center.
- HPE Cray Supercomputing EX254n accelerator blade โ Ideal for AI and machine learning (ML) workloads, this will be one of the first supercomputing blades to use the NVIDIA Grace Hopper GH200 Superchip. The HPE Cray EX254n is expected to become available this winter.
- HPE Cray Supercomputing EX255a accelerator blade โ Ideal for AI and HPC workloads, this will be the first supercomputing accelerator blade to feature the worldโs first APU accelerator for AI and HPC, the AMD Instinctโข MI300A APU. The HPE Cray Supercomputing EX255a accelerator blade will become available in early 2024 and will power the Lawrence Livermore National Laboratoryโs El Capitan Supercomputer.
Accelerate Generative AI with new turnkey solution
HPE also announced a new supercomputing solution for generative AI, which combines industry-leading supercomputing hardware with a full software toolset, data fabric and services to provide a simplified, integrated solution specifically designed for high impact AI outcomes. This turnkey solution packages HPE Cray Supercomputing EX2500 with the NVIDIA Grace Hopper GH200 Superchip, HPE Cray Programming Environment, HPE Machine Learning Development Environment and HPE Complete Care services to help customers more efficiently scale their AI workloads with sustainable direct liquid cooling and full-stack support. Read the full supercomputing solution for generative AI press release here.
New analytics software provides insights into unique workload performance and resource utilization
A strong network bolsters the incredible speed of supercomputers. For that reason, HPE announced HPE Cray Supercomputing Fabric AIOps Software designed to provide unparalleled insights into HPC and AI infrastructure. This offering will help HPE Slingshot customers maximize performance, reduce downtime and gain highly granular insight into workload fabric resource utilization.
HPE has shipped well over one million HPE Slingshot switch ports to date. The release of HPE Cray Slingshot Fabric AIOps Software offers a valuable addition to the diagnostics and analytics capabilities available today. This software represents a significant milestone to the ever-expanding HPE Slingshot product offering and will be available in early 2024.
Whether youโre driving innovation in national defense, manufacturing, health and life sciences or another discipline, HPEโs supercomputing team is committed to partnering with you to accelerate innovation and discovery in the AI era in a sustainable way.
[i] HPE Cray is the industry-leader per Hyperion Research, HPC Market Update, April 2023
Server Experts
Hewlett Packard Enterprise
twitter.com/HPE_HPC
linkedin.com/showcase/hpe-servers-and-systems/
hpe.com/servers
ComputeExperts
Our team of Hewlett Packard Enterprise server experts helps you to dive deep into relevant infrastructure topics.