Hewlett Packard Enterprise/SGI Cheyenne and Pleiades systems named on TOP500 list
14 Nov 2016 Palo Alto - The NASA Ames Research Center's supercomputer, Pleiades, was ranked #13 and is the most powerful HPE/SGI supercomputer on the prestigious TOP500 list at the SC16 Conference, the international conference for High-Performance Computing (HPC), Networking, Storage and Analysis. In addition, the National Center for Atmospheric Research's (NCAR) SGI ICE XA supercomputer, Cheyenne, was ranked #20 in the same list.
NASA's Ames Research Center's Pleaides helps the space agency to pioneer the future of space exploration, scientific discovery and aeronautics research through increased computational capabilities. The system supports NASA's real-world engineering applications and also improves modelling and simulation methods for developing aerospace technologies.
NCAR's Cheyenne is one of the world's most powerful commercial supercomputers with 5.33 petaflops. Cheyenne is built on SGI's flagship SGI ICE XA system which enables scientists, researchers and engineers to accelerate computational breakthroughs and provide more actionable projections about the impact of climate change for specific regions and assist agencies throughout the world to develop more accurate weather predictions on a local and global scale.
On November 1, 2016, HPE announced the completion of the acquisition of SGI. HPE and SGI combined hold the most number of systems, 140, recognized in the TOP500 list, showcasing innovation and HPC leadership. Cheyenne joins the already impressive list of HPE and SGI systems on the TOP500. This continued progression reflects HPE/SGI's joint commitment to the HPC market and its customers.
"HPC is rapidly becoming one of the essential components for the digital economy across organisations of all types and sizes", stated Bill Mannel, vice president and general manager, HPC segment solutions, HPE. "Once again, we are excited to be recognized on the TOP500 list as having the most supercomputers. The addition of SGI solidifies HPEs position of having 5 systems of the Top 50 as well."
"It's always rewarding to see our customer's systems recognized in the industry's most respected index, the TOP500 list", stated Gabriel Broner, vice president and general manager, high-performance computing at SGI. "NCAR understands the benefits of investing in supercomputing and together we've delivered a system that will offer new modelling capabilities to provide groundbreaking research insights in a fraction of the time."
HPE and SGI also showcased new technologies at SC16, including:
- The Multi-Rail Lustre capability from HPE/SGI and Intel which delivers highly scalable data pipelines by bonding the throughput of multiple Intel Omni-Path or InfiniBand connections to support the rapid movement of large data sets for HPC and data analytics applications. SGI demonstrated the SGI UV 300 as a large memory data analytics and computation node with Multi-Rail Lustre, Intel NVMe storage utilizing Intel Omni-Path interconnect.
- HPE/SGI Management Suite and HPE/SGI Performance Suite now fully support SGI's Intel Xeon Phi processor-based platforms as well as Intel Omni-Path. SGI Management Suite now also includes an advanced analytical power management feature which helps customers to predict and plan their power usage more easily by analyzing up to three months of historical power usage data.
- An HPE/SGI OpenStack Reference Architecture is now available and provides guidance for integrating SGI Rackable scale-out servers and optional SGI UV scale-up servers with SUSE OpenStack Cloud software. Following these recommendations and best practices allows organisations to quickly and confidently deploy the required infrastructure that yields a secure, scalable, performant and highly available private Cloud instance for HPC workloads, including testing, development or production usage scenarios.
- HPE/SGI demonstrated how computational fluid dynamics (CFD) enhances aerodynamic design in the Elemental Car. London Computational Solutions used SGI's HPC capabilities to provide more downforce for Elemental Car's top rated racing vehicle. Using CFD powered by SGI, the car was redesigned for improved aerodynamics in just four weeks vs. what previously took nine months.
- The HPE Quantitative Finance Library solution designed for financial services customers leverages the Altimesh tool with the HPE Apollo 6000 system to enhance application software performance through multiple levels of parallelism. This will increase business productivity through automation and accelerate applications leading to more competitive differentiation in high-frequency trading.
- The HPE Next Generation Sequencing solution for life science customers provides optimized, flexible reference configurations for HPE Apollo 2000 to enable faster genome assembly and sequencing while lowering total cost of ownership resulting in increased revenue for bio-sciences organisations.
- The HPE High Performance Remote Visualization for Computer Aided Engineering (CAE) delivers 84 percent lower hardware and software acquisition costs, 45 frames per second and up to 32 users per GPU while continuing to offer the best performance experience. The solution, featuring HPE Apollo 2000 servers delivers the ideal solution for engineers and designers seeking robust remote visualization capabilities.
- The new version 2.0 of the HPE Core HPC Software Stack combines the capabilities of the new version of HPE Insight Cluster Management Utility (CMU) v8.1 with OpenHPC application compatibility and multiple partner component integrations. CMU v8.1 now provides limitless HPC cluster scalability with faster provisioning along with more comprehensive monitoring and management to deliver faster time to value - from bare metal to fully functional cluster in two hours. Free downloads of this entire stack for up to 32 nodes and a 120 day trial license for larger clusters are now available. In addition, a HPE supported version of this software stack will be available in the first calendar quarter of 2017.