AUSTIN, Tex., Nov. 16 — The Stephen Hawking Centre for Theoretical Cosmology (CTC) at the University of Cambridge has been recognized in the annual HPCwire Readers’ and Editors’ Choice Awards, with the following honour:
- Readers Choice: Best Use of High Performance Data Analytics
CTC operates the COSMOS hybrid shared-memory supercomputer (www.cosmos.damtp.cam.ac.uk) which is part of the STFC DiRAC HPC Facility and they are also an Intel Parallel Computing Centre (IPCC). The award was for the impressive many-core acceleration of the MODAL analysis pipeline which offered new statistical insights from the Cosmic Microwave Background as observed by the ESA Planck Satellite – Planck is also an STFC-funded project. The work was achieved on the first Intel Xeon Phi-enabled SGI UV2000 with its co-designed ‘MG Blade’ Phi-housing, part of a portfolio of innovative
The Readers Choice award was presented at the 2015 International Conference for High Performance Computing, Networking, Storage and Analysis (SC15), in
Professor Paul Shellard, CTC Director, said: “We are thrilled at the Centre for Theoretical Cosmology and COSMOS IPCC to have received this international award in high performance computing. It is recognition of a unique synergy that we have developed between world-leading researchers from the STFC DiRAC HPC Facility and industry-leading vendors like Intel and SGI which aims to get maximum impact from new many-core technologies for our data analytic pipelines. Dramatic speed-ups have been achieved for our Planck satellite analysis and other codes through a potent combination of new parallel programming paradigms and architectural co-design; these capabilities are opening up new windows on our Universe.”
For many years, STFC-funded scientists in
Center focusing on Xeon Phi porting and optimization efforts on their unique hybrid UV2000 system co-designed for many-core acceleration with SGI. This IPCC support coincided with Cosmic Microwave Background (CMB) measurements by the Planck satellite which provided the first high resolution temperature and polarization maps of the entire sky. The COSMOS IPCC team adapted the main workhorse non-Gaussian statistical correlation code, MODAL, which is designed to analyze very small CMB fluctuations in the Planck data; it aims to provide insight into new physics theories about how structures formed in the Universe. This is a computationally daunting task and a complete analysis for three-point correlations would have taken unfeasibly long to perform even on the largest supercomputers available to the researchers. The use of the hybrid UV2000 + Xeon Phi system, combined with the optimization and modernization effort of the COSMOS IPCC team, resulted in runtimes being cut by a factor 1/100-1/1000 which meant we could meet the tight ESA timescales available for the analysis.
DiRAC is the integrated supercomputing facility for HPC-based research in particle physics, astronomy and cosmology, areas in which the
Tom Tabor, CEO of Tabor Communications, publisher of HPCwire, said, “HPCwire readers are among the most informed in the HPC community and these awards are ultimately given to the organizations that are making the greatest impact in advancing technology and humanity itself through high performance computing. The HPCwire Readers’ and Editors’ Choice Awards send a strong message of support and appreciation from those in the global HPC community. We are proud to be able to recognize these efforts each your and our congratulations go out to all the winners.”
About The Stephen Hawking Centre for Theoretical Cosmology
The Centre for Theoretical Cosmology (CTC) was established by Professor Stephen Hawking in 2007 at the Department of Applied Mathematics and Theoretical Physics in the
About HPCwire
HPCwire is the #1 news and information resource covering the fastest computers in the world and the people who run them. With a legacy dating back to 1986, HPCwire has enjoyed a legacy of world-class editorial and journalism, making it the news source of choice selected by science, technology and business professionals interested in high performance and data-intensive computing. Visit HPCwire at www.hpcwire.com.
No hay comentarios:
Publicar un comentario