Supercomputing
The term supercomputing refers to efficiently solving massively complex or data-intensive problems using the concentrated compute resources of multiple computer systems working in parallel. ComnetCo and HPE can deliver everything from custom scale-up and scale-out architectures to fully burned-in and ready-to-run petascale systems like HPE Cray supercomputers.
Computing Power Reimagined – Supercomputing
Supercomputing completely rethinks how technology comes together to conquer the greatest challenges of today and prepare for tomorrow’s even bigger ones. However, the term supercomputing is often misconstrued, causing organizations to shy away from high-performance computing (HPC) opportunities.
Many organizations believe they lack the financial resources, expertise, or even the need to adopt the latest technologies. As a result, a number of markets and industries have left the benefits of supercomputing widely underutilized. Let ComnetCo help you to demystify supercomputing and apply its groundbreaking applications to your toughest, most time-sensitive science and engineering jobs. With our consulting services you will find it is much easier than you think to speed your time-to-discovery, time-to-insights, and time-to-market—which all translates into better ROI.
As data and workloads become too complex for traditional compute infrastructures to handle, supercomputing technologies can deliver greater compute power and speed to overcome the limitations of inflexible architectures. Consequently, organizations can optimize an extensive range of tasks—from small-scale activities to high-end workloads. These capabilities can make sense out of events that are too difficult to see or measure in real life. The data-intensive nature of new application requirements will also require ongoing innovation for the foreseeable future. Supercomputing provides the compute capabilities to accelerate these data-driven workloads.
Advancements in supercomputing technologies are changing the face of the modern data center, as organizations implement capabilities today that are relevant for years to come. Addressing important scientific, engineering, and business challenges depends on continued investments in supercomputing. Strategic adoption of these technologies gives organizations the potential to evolve and lead their industries.
Virtually every industry and research organization can benefit from supercomputing technologies, and ComnetCo, HPE, and NVIDIA are working to bring these capabilities into the data center. Through a wide variety of use cases, numerous organizations are realizing better outcomes, whether that means increased profits and productivity, delivering better services, or discovering a breakthrough. Supercomputing technologies have impacted operations in sectors including manufacturing, energy and gas, financial services, healthcare and life sciences, earth sciences and climate, and academia.
HPE Cray EX
Performs Like a Supercomputer, Runs Like a Cloud
The HPE Cray family of supercomputers provides Scientific & Research Computing Centers with complete solutions that are tightly integrated and performance-tuned at the factory. These iconic supercomputers offer the best system performance while delivering customers a new standard in flexibility, manageability, and resiliency.
A number of challenges emerge or grow when core counts increase, compute node architectures proliferate, and workflows expand to accommodate production AI at scale. HPE Cray supercomputers incorporate leading-edge hardware and software innovations that work to improve system bottlenecks, manageability, and job completion rates. By eliminating the distinction between clusters and supercomputers with a single new system architecture, they also enable a choice of computational infrastructure without having to compromise.
The latest HPE Cray EX system hardware architecture is designed to support the wide array of processor and accelerator options—those both available today and in the future. The flexible architecture supports CPUs, GPUs and fabric interconnects in both liquid- and air-cooled cabinetry to meet end-user requirements for density, performance, and efficiency. For customers requiring the greatest performance, density, and efficiency for large-scale systems, the HPE Cray EX supercomputer is available in liquid-cooled cabinetry, which supports Direct Liquid Cooling (DLC) to all components in a compact, bladed configuration.
The HPE Cray EX Supercomputer solution includes an optimized HPC interconnect—HPE Slingshot—which delivers high-throughput and sustained performance at Exascale. Slingshot offers the ability to run diverse workloads simultaneously without interference plus the flexibility to enable new approaches to computing such as supercomputing via the cloud.
HPE Slingshot
An Innovative Interconnect for the Next Generation of HPC
HPE Cray Slingshot offers a modern, high-performance interconnect for HPC and AI clusters that delivers industry-leading performance, bandwidth, and low latency for HPC, AI/ML, and data analytics applications. Cray’s 8th generation interconnect architecture includes powerful innovations engineered from the silicon up. These advances combine the best of a high-performance supercomputing-class interconnect with Ethernet compatibility. They also enable a unified high-performing cluster that delivers superior performance on both traditional modeling and simulation codes alongside native sockets-based applications. Additionally, they provide direct Ethernet connectivity to storage without requiring gateway nodes.
Beyond raw speed (25.6 Tbps of bidirectional bandwidth per switch and 200 Gbps endpoints), HPE Slingshot dramatically reduces loaded latency—the true determinant of realized performance on tightly coupled HPC and AI workloads. Innovative congestion management further reduces tail latency and run-to-run execution time variability. Fine-grained adaptive routing also ensures high utilization of available bandwidth.
With HPE Slingshot for HPC clusters and supercomputers, the same interconnect in the world’s most powerful Exascale systems can now be deployed across a range of HPC system sizes and configurations based on HPE Apollo and HPE ProLiant compute nodes.
HPE Slingshot Highlights
- High bandwidth and low latency
- Automatic congestion management
- Fine-grained adaptive routing
- Superior Quality-of-Service (QoS)
- Low-diameter network
- Ethernet standards
Cray Software
An Essential Software Suite for HPC Organizations Developing Their Own Code
As HPC systems grow larger and more complex—and feature—different heterogeneous hardware components to accommodate different workloads, programmers need different software development tools. A tool-rich environment can ease application development and aid with targeted performance tuning and optimization.
HPE Cray Programming Environment suite offers just that—a comprehensive set of tools for developing, porting, debugging, and tuning of applications. The environment provides:
- Compilers, programming languages, and models
- Scientific, math, and I/O libraries
- Debugging tools
- Profiling and performance optimization tools
