HPE Cray Supercomputers

  1. Home
  2. >
  3. Solutions
  4. >
  5. HPE Cray Supercomputers

HPE Cray supercomputers offer turnkey solutions for the Next Era of Supercomputing. Fully burned-in and configured at the factory, they represent an entirely new design—created from the ground up to handle today’s massive converged modeling, simulation, AI, and analytics

HPE Cray EX

For the last 20+ years, most supercomputing solutions have shared a similar design—a large scale-out cluster of servers based on an industry-leading microprocessor/GPU architecture and connected on an industry-leading microprocessor/GPU architecture and connected by an industry-standard interconnect. This approach has worked well, providing the highest performance required for challenges in scientific and engineering fields such as weather and climate, computational fluid dynamics, mechanical design, security, defense applications, and more. 

Nonetheless, the next decade promises to bring a new set of workloads and a dizzying array of potential processing solutions that challenge the high-performance computing (HPC) status quo. Relentless data growth coupled with a global business imperative for digital transformation is driving more and bigger workloads. Traditional modeling and simulation workloads are fusing with AI, analytics, and the Internet of Things (IoT) to create massive business-critical workflows. 

At the same time, multiple x86, Arm, GPU, and field-programmable gate array (FPGA) vendors are expected to deliver compelling offerings. This makes it increasingly difficult to predict which future core processing architectures will offer and provide the best value. It means HPC users will need systems that can: 

  • Support a variety of compute platforms 
  • Upgrade to new architectures as they become available 

This confluence of factors has forced a complete rethinking of supercomputing architectures. It also signals a major technology inflection point and a new era for HPC. More than a simple speed milestone, the exascale era is a new set of capabilities for a new set of workloads that is transforming every field of inquiry. 

The HPE Cray supercomputing architecture is the HPE answer to these new demands. It comprises an entirely new design in which the compute, interconnect, software, and storage have been rethought and rearchitected to meet current and future system requirements across HPC, AI, and converged workloads. HPE Cray EX is designed to run data-intensive and diverse workloads in parallel. Hardware and software innovations further limit system bottlenecks in processing, data movement, and I/O.  At the same time, the architecture provides greater flexibility for multiple processor and accelerator architectures as well as a dynamic choice of system interconnect technologies.