From the formation of black holes to the origins of space, data is the key to unlocking the secrets of the early universe. The challenge is, scientists must analyze 14 billion years’ worth of information to reach these conclusions. This herculean task requires a compute platform that can utilize massive datasets to harness groundbreaking insights.
Hewlett Packard Enterprise (HPE) has joined forces with the Faculty of Mathematics at the University of Cambridge to accelerate computational mathematics and fuel scientific discovery. In a powerful partnership with the Centre for Theoretical Cosmology (COSMOS), HPE and distinguished Cambridge professor and COSMOS founder, Stephen Hawking, hope to answer some of man’s greatest questions and extend human knowledge beyond time and space.
Driving scientific research with HPC
Hawking founded the COSMOS supercomputer facility in 1997, in search of an “ultimate theory” which would enable scientists to predict “everything in the universe.” Today, this collaboration has introduced a new high performance computing (HPC) machine designed to collect and analyze a deluge of complex data. A new generation of HPC solutions from HPE will help scientists identify and study pivotal events in our universe—like the Big Bang—as well as drive progress in cosmology and relativity.
The latest COSMOS supercomputer combines the HPE Superdome Flex, an in-memory compute platform, with the high density and extreme performance of HPE Apollo servers. And backed by a highly efficient Intel® Xeon Phi™ system, researchers can leverage a variety of data, from gravitational waves and cosmic microwaves, to the distribution of stars and galaxies. This robust architecture can support research and exploration into a wide range of disciplines, such as environmental sciences and medical imaging, enabling scientists to identify critical signatures in rising volumes of data.
Accelerating insight with memory-driven computing
To keep pace with escalating data requirements, HPE is leveraging memory-driven computing to deliver extreme scalability and performance. Memory-driven computing puts memory at the center of the architecture, instead of the processor. This dramatically increases efficiencies to process data faster and turbocharge data analytics, from the core to the intelligent edge.
The HPE Superdome Flex is the world’s most scalable and modular in-memory computing platform. Storing data in memory allows scientists to perform analytics almost instantly, achieving actionable insights in real time. With a large, scalable core of memory, the system can scale from 4 to 32 sockets and 768GB to 48TB of shared memory in a single HPC system. The HPE Superdome Flex delivers unmatched compute power to streamline the most demanding workloads, and because it is cloud-ready, organizations can easily scale up or out to solve data-intensive problems.
COSMOS scientists are benefiting immensely from these next-generation capabilities:
- Ultra-low latency and high bandwidth empower breakthrough performance at scale
- 45% lower acquisition cost over the previous generation
- Industry-leading reliability, accessibility, and serviceability (RAS), resulting in 99.999% availability
- Error Analysis Engine predicts hardware faults and reduces human error without operator assistance
- HPE’s “firmware first” approach contains critical errors before an interruption can occur
According to Randy Meyer, VP and GM, Synergy and Mission Critical Servers for HPE, “The in-memory computing capability of HPE Superdome Flex is uniquely suited to meet the needs of the COSMOS research group. The platform will enable the research team to analyze huge datasets and in real time.” The COSMOS supercomputer has the capacity to support mission-critical workloads in addition to data-centric research in a variety of mathematical fields, enabling more than 50 scientists to pursue revolutionary intelligence.
Memory-driven computing is revolutionizing compute capabilities and pioneering a new era of scientific discovery. To learn more about next-generation HPC servers, I invite you to follow me on Twitter at @Bill_Mannel. And for more information on collaborative partnerships that are transforming HPC, check out @HPE_HPC.