Collaboration and Innovation: SDSC Launches Petascale ‘Comet’ Supercomputer

Published October 23, 2015

When the San Diego Supercomputer Center (SDSC) launched its first supercomputer, a Cray XMP-48 in late 1985, it was about as powerful as an iPhone is today. Now SDSC has formally taken the wraps off Comet, a new petascale supercomputer that is over 2 million times more powerful than that first system. 

With the ability to perform almost two million billion operations or calculations per second, Comet is designed to transform scientific research by expanding computational access to a larger number of researchers working across a wider range of domains.

“The San Diego Supercomputer Center plays a vital role in fulfilling our vision to solve our world’s most pressing research challenges,” UC San Diego Chancellor Pradeep K. Khosla told a capacity audience at SDSC last week as the Center showcased Comet as part of its 30^th anniversary celebration.

Chancellor Khosla spoke about how SDSC has become a national leader in cyberinfrastructure, providing the high-performance computing, storage, networking, and expertise needed to harness the university’s collective research efforts efficiently and seamlessly. He also emphasized SDSC’s role in advancing UC San Diego’s Strategic Plan goals and research endeavors.

“Today, UC San Diego is a $1 billion research enterprise,” noted Khosla, adding that “SDSC has been a trail blazer for academic computing.”

“We are fortunate that SDSC has been one of the national leaders – research-oriented pioneers – in building an advanced cyberinfrastructure ‘nervous system’ for the academic and scientific communities,” said UC San Diego Vice Chancellor for Research Sandra A. Brown, another featured speaker at the event.

The result of a National Science Foundation grant valued at almost $24 million including hardware and operating funds, Comet is designed to meet the needs of what is often referred to as the ‘long tail’ of science – the idea that the large number of modest-sized computationally-based research projects represent, in aggregate, a tremendous amount of research that can yield scientific advances and discovery. A video about Comet can be viewed below.

Comet joins SDSC’s Gordon supercomputer as another key resource within the NSF’s XSEDE (eXtreme Science and Engineering Discovery Environment) program, which comprises the most advanced collection of integrated digital resources and services in the world for use by U.S.-based academic researchers.

“The launch of Comet marks yet another stage in SDSC’s leadership in the national cyberinfrastructure ecosystem,” said James Kurose, Assistant Director of the NSF’s Computer and Information Science and Engineering (CISE) Directorate, in remarks at the SDSC event. “Through this launch and the extraordinary computing capabilities of SDSC, the center will continue to expand the frontiers of science and engineering, allowing researchers to open new windows into phenomena as vast as the Universe and as small as nanoparticles.”

SDSC Director Michael Norman noted that SDSC’s mission has expanded over its three decades to encompass much more than advanced computation, including a host of services related to the voluminous amount of digitally based information generated daily, and systems designed to analyze, store, and share that data.

“In recent years the research community has moved into a new era of scientific endeavor based on computational science, now best described as data-intensive science,” said Norman. “The term ‘big data’ became the short-hand description for this, or, for academia, ‘data science and engineering.’ This convergence of computational science with data science and engineering rests on an inherent reliance of interdisciplinary collaborations, which is needed to solve the grand research challenges of our times.”

“Comet’s innovative design makes it ideal for supporting a broad range of research and computing modalities.  Two distinctive features – support for science gateways and high-performance virtualization – will significantly expand the community of researchers with access to high-performance computing resources,” said SDSC Deputy Director Shawn Strande, who also is Comet’s program manager. “Comet is expected to reach an active research community of over 10,000 users, and is destined to become one of the most productive HPC systems available to the academic research community.”

In addition to highlighting SDSC’s major milestones in an expanded timeline covering its 30-year history, the Center also announced the launch of a new fund-raising campaign for UC San Diego, in partnership with SDSC, designed to tackle some of the grand research challenges facing the State of California and beyond outlined previously in the campus’ Strategic Plan. Details about this new effort will be announced at a later date.

How SDSC’s ‘Comet’ Supercomputer is Serving Science and Society Comet is configured to help transform advanced computing by expanding access and capacity not only among research domains that typically rely on high-performance computing – such as chemistry and biophysics – but among domains which are relatively new to using  supercomputers, such as genomics, finance, and the social sciences. Some of the domains already being served by Comet include: Astrophysics: Supercomputers can greatly accelerate timescales for researching the origins of the universe. Neurosciences, Brain Research: SDSC’s Neuroscience Gateways project will contribute to the national BRAIN initiative announced by the Obama Administration to deepen our understanding of the human brain. Social Sciences: Sociologists and political scientists are analyzing newly accessible data sets to study censorship of the press, factors that affect participation in the political process, and the properties of social networks. Molecular Science: Studying the properties of lipids, proteins, nucleic acids, and small molecules can advance our understanding of biophysical processes at the atomic scale, leading to new drug designs and reducing disease. DNA Nanostructures: Conducting nanoscale biomolecular research could lead to low-cost DNA sequencing technologies, and in turn create targeted drug delivery systems and help explain the molecular causes of disease. Alternative Energy Solutions/New Materials Research: Finding new and more efficient solutions to energy harvesting, nanoporous membranes for water desalinization, solar thermal fuels, and more. Fluid Turbulent Physics: Supercomputers can create highly detailed simulations to track ocean currents or improve industry methods related to the discharge of pollutants, or oil flow in pipelines. Climate Change/Environmental Sciences: Modeling atmospheric aerosols, identified as influencing the chemical composition and radiative balance of the troposphere, has direct implications for our climate and public health. Seismic Research/Disaster Prevention: Keys to hazard management for major earthquakes, hurricanes, and wildfires include the ability to predict a wide range of possibilities. Supercomputer-generated simulations are used to inform decision-making strategies. The Tree of Life: Biologists construct phylogenetic trees to capture the evolutionary relationship between species, and help us better understand the functions and interactions of genes, the origin and spread of diseases, the co-evolution of hosts and parasites, and migration of human populations.  Key Features of Comet ~2 petaflops of overall peak performance – one million billion operations or calculations per second. Dell compute nodes using next-generation Intel Xeon processors, 27 racks of compute nodes totaling 1,944 nodes or 46,656 cores. 128 GB (gigabytes) of DRAM and 320 GB (gigabytes) of flash memory per standard compute node 72 nodes per rack with full bisection InfiniBand FDR interconnect in each rack, and a 4:1 bisection cross-rack interconnect Additional GPU and large-memory (1.5 Terabytes) nodes for applications such as visualization, molecular dynamics simulations, or de novo genome assembly 7 PB (petabytes) of Lustre-based high-performance storage from Aeon, and 6 PB of durable storage for data reliability First XSEDE production system to support high-performance virtualization

About SDSC

As an Organized Research Unit of UC San Diego, SDSC is considered a leader in data-intensive computing and cyberinfrastructure, providing resources, services, and expertise to the national research community, including industry and academia. Cyberinfrastructure refers to an accessible, integrated network of computer-based resources and expertise, focused on accelerating scientific inquiry and discovery. SDSC supports hundreds of multidisciplinary programs spanning a wide variety of domains, from earth sciences and biology to astrophysics, bioinformatics, and health IT. SDSC’s Comet joins the Center’s data-intensive Gordon cluster. SDSC is a partner in XSEDE (eXtreme Science and Engineering Discovery Environment), the most advanced collection of integrated digital resources and services in the world.