News

Flux sees increases in users, jobs, departments

Posted on by

The Flux high-performance computing cluster saw a 43 percent increase in number of jobs between fiscal 2011 and 2012.

About 1.82 million jobs were run on Flux in fiscal 2012, compared to 1.27 million in 2011.

Flux, provided by ARC, also had a 49 percent increase in users (from 482 to 718) and a 31 percent increase in departments with researchers using the service (from 45 to 59).

“The rising numbers we’ve seen in users, jobs, and departments using Flux reflect the increasingly crucial role that high-performance computing plays in research across fields,” said Ken Powell, Interim Associate Vice President for Research – Cyberinfrastructure.

Most users of Flux come from LSA, the Medical School, and the College of Engineering. Those three academic units, along with OVPR and the Provost, were the original funders of the service.

But a growing number of users come from outside those disciplines as well.

For example, David Bieri, an assistant professor of Urban Planning in the Taubman College of Architecture and Urban Planning, and his colleagues had a research question that needed an extraordinary amount of computing power to address: How much do consumers spend on local amenities, as reflected in housing prices and wages? The answer was contained in massive datasets that could not be handled by a standard computer.

“Our work involves huge amounts of microdata from the census and other sources … we had a sample that contained 30 million observations and over 2,000 variables,” Bieri said. “We were quickly outpacing the computational abilities of a laptop or even the most expensive desktop computer.”

Bieri used Flux and the multiprocessor version of Stata, which is included in the cost of an allocation, to analyze the data.

Bieri also relied the support offered by ARC and CoE — from seminars introducing Flux to Web resources to in-person help from CAEN HPC staff member Bennet Fauber. CAEN operates Flux and provides user support.

Bieri’s findings, that spending on local amenities varies greatly across the country, and can be as high as 25 percent of household income in some places, are being reviewed for publication by an applied economics journal.

Keynote speakers set for CI Days 2013

Posted on by

Abby Smith Rumsey, a historian of ideas who is an expert in digital preservation, and Lada Adamic, an associate professor at the U-M School of Information and Center for the Study of Complex Systems who is working at Facebook while on leave from the University, will deliver the keynote addresses at Cyberinfrastructure (CI) Days 2013.

CI Days, the annual event for the U-M advanced computing research community organized by ARC, will take place Nov. 13-14 at the Rackham Graduate School on U-M’s Central Campus.

Rumsey has been working with the National Digital Information Infrastructure and Preservation Program at the Library of Congress to develop a national strategy to identify, collect, and preserve digital content of long-term value.

Adamic’s research focuses on the structure and dynamics of social and information networks, with an emphasis on information diffusion, expertise sharing, and online communities.

Report proposes creation of “visualization hubs” on U-M campus

Posted on by

The U-M Future of Visualization Committee issued a report June 4 calling for creation of two “visualization hubs,” which would serve to make computing visualization services more accessible to the campus research community.

The committee was created by Dan Atkins with the charge of evaluating existing visualization technologies and methods on campus; developing an action plan for addressing deficiencies in visualization needs; establishing a group of visualization leaders; and communicating with the community on visualization topics. It is composed of faculty members and staff from ARC, University Libraries, Dentistry, LSA, the Medical School, ITS, Architecture and Urban Planning, Atmospheric and Oceanic and Space Sciences, and the College of Engineering.

The hubs envisioned by the committee would leverage existing resources and consist of advanced workstations, high bandwidth connectivity, and collaborative learning spaces, with a support model based on that of the UM3D Lab and Flux. The hardware and software would be configured to allow departments or individuals to purchase their own resources in a way that would reduce fragmentation and allow for efficient support, training, and maintenance.

A full copy of the report is available here: Report_VisualizationOnCampus_2013.

Job listing: University of Kentucky visualization center director

Posted on by

The University of Kentucky is seeking candidates for director of its Center for Visualization & Virtual Environments. Candidates with expertise in any area related to visualization, including but not limited to machine vision, medical imaging, and pattern recognition, will be considered for this tenured faculty position.

See http://vis.uky.edu/blog/2013/04/23/vis-center-director-search/ for more information.

Exascale computing: Coming soon?

Posted on by

The potential of exascale computing — defined as high-performance computers capable of executing one quintillion floating point operations (or FLOPS) per second — has received a lot of attention over the past several weeks. Exascale computing would be 1,000 times faster than the current petaflop computers, and holds the promise of “vastly more accurate predictive models and the analysis of massive quantities of data, producing quantum advances” in fields like climate, transportation, and energy, according to a 2010 Department of Energy report.

On May 22, a U.S. House subcommittee held a hearing titled “America’s Next Generation Supercomputer: The Exascale Challenge,” with the stated goals of examining the challenges and opportunities of exascale computing, and reviewing draft legislation directing the Department of Energy to develop an exascale computing system.

To read the testimony and watch an archived webcast of the hearing, visit http://science.house.gov/hearing/subcommittee-energy-hearing-exascale-computing-challenges-and-opportunities

Horst Simon, Deputy Director of the Lawrence Berkeley National Laboratory, also made news with his prediction that exascale computing won’t be achieved by 2020 (see http://www.hpcwire.com/hpcwire/2013-05-15/no_exascale_for_you_an_interview_with_nersc_s_horst_simon.html), and his willingness to bet $2,000 of his own money to back up that prediction (see http://www.extremetech.com/computing/155941-supercomputing-director-bets-2000-that-we-wont-have-exascale-computing-by-2020).

A podcast from Nicole Hemsoth, Editor of HPCwire, and Addison Snell, CEO of InterSect360 Research, posted at HPCWire, discusses the congressional hearing: http://www.hpcwire.com/hpcwire/2013-05-24/podcast:_exascale_equals_no_nukes_ibm_says_lets_build_a_smarter_hpc.html.

Job listing: Assistant Director for Research Cyberinfrastructure, University of Minnesota

Posted on by

The Minnesota Supercomputing Institute at the University of Minnesota is seeking an Assistant Director for Research Cyberinfrastructure.

The listing seeks candidates with particular experience and expertise in data intensive research, and sound knowledge and demostrable abilities in high-performance computing hardware platforms,scientific software environments, and disciplinary research in a data intensive field.

For more information, see https://www.msi.umn.edu/content/assistant-director-research-cyberinfrastructure and for information on applying, visit https://employment.umn.edu/applicants/Central?quickFind=111219.

ORCI is now ARC

Featured

Posted on by

To clarify our mission, ORCI will now be known as Advanced Research Computing at U-M, or ARC. This name reflects our goal of providing high-performance computing resources and support for the U-M advanced research computing community.

ARC’s mission is to:

  • Ensure the availability and accessibility of advanced computing resources
  • Assess researchers’ computing needs and connect you to the appropriate resources on an ongoing basis
  • Help researchers explore the potential of advanced computing
  • Promote collaboration in proposal writing, mentoring, and best practices relating to advanced research computing
  • Inform the research community of the latest developments and capabilities in advanced computing.

ARC will continue to operate under the Office of the Vice President for Research. We will continue to be the provider of the Flux cluster and organizer of the annual CI Days event, as well as other initiatives and programs to enhance advanced computing resources on campus.

INCITE call for proposals is open — April 15 through June 28

Posted on by

The 2014 Innovative and Novel Computational Impact on Theory and Experiment (INCITE) call for proposals is open from April 15 through June 28. The program enables transformational advances in science and technology for computationally intensive, large-scale research projects through large allocations of computer time and supporting resources at the Argonne and Oak Ridge Leadership Computing Facility (LCF) centers, operated by the U.S. Department of Energy Office of Science.

INCITE seeks research enterprises for capability computing: production simulations—including ensembles—that use a large fraction of the LCF systems or require the unique LCF architectural infrastructure for high-impact projects that cannot be performed anywhere else.

INCITE is currently soliciting proposals of research for awards of time on the 27-petaflop Cray XK7 “Titan” and the 10-petaflop IBM Blue Gene/Q “Mira” beginning in calendar year (CY) 2014. More than 5 billion core hours will be allocated for CY 2014. Average awards per project are expected to be on the order of 50 million core-hours for Titan and 100 million core hours for Mira but could be as much as several hundred million core hours. Proposals may be for up to 3 years.

For more information on the INCITE program and links to the proposal form, see http://www.doeleadershipcomputing.org/. You may address specific questions to the INCITE Manager, Julia White, at incite@doeleadershipcomputing.org.

Request for Information

We invite prospective INCITE proposal authors to respond to a Request for Information to inform INCITE management of expected proposal topics. See http://proposals.doeleadershipcomputing.org/allocations/incite/. You are encouraged to respond if you are contemplating submitting an INCITE proposal. The information requested is not a prerequisite for proposal submittal, nor will it limit any requests you may decide to make in your INCITE proposal.

INCITE Proposal Writing Webinars April 25th and May 14th

To help you prepare an INCITE proposal or learn more about the program, two INCITE Proposal Writing Webinars will be offered on April 25, 2013, at 4 p.m. EDT and on May 14, 2013, at 10 a.m. EDT. The 1.5‑hour webinars will also include discussion on the Director’s Discretionary Program, a way to request early access to port, tune, and scale your codes in preparation for an INCITE application. For more details and early registration for these events, seehttp://www.alcf.anl.gov/workshops/incite-2014-webinars for the April 25th event or http://www.olcf.ornl.gov/training-event/incite-proposal-writing-webinar/ for the May 14th event. See http://www.doeleadershipcomputing.org for additional event information.

XSEDE upgrades network to Internet2′s 100G network

Posted on by

Internet2, operator of the nation’s fastest research and education network, and XSEDE, the Extreme Science and Engineering Discovery Environment, announced April 10 that XSEDE has migrated and upgraded its network backbone infrastructure to the Internet2 Network. The upgrade allows XSEDE to use Internet2’s new 100G (Gigabit Ethernet)-enabled and 8.8-terabit-per-second optical network, platform, services and technologies.

XSEDE is a National Science Foundation-supported project that brings together 17 supercomputers, visualization and data analysis engines, and data storage resources along with data collections, computational tools, and services to support science and engineering research and education across the United States. More than 8,000 scientists regularly use XSEDE to complete thousands of research projects supporting more than 2,000 publications annually.

“Advanced networking is critical to XSEDE in order to support the researchers and educators who are making innovative use of our computing, data, and visualization resources that are distributed across the country,” said XSEDE project director John Towns. “Our partnership with Internet2 enables us to continue to provide advanced, powerful, robust services for discovery and innovation.”

Internet2 and the XSEDEnet networking group will work together to configure a private network between the XSEDE service provider sites across the shared backbone. A majority of sites will connect to the nearest Internet2 Advanced Layer 2 Service node at 10G and share bandwidth across a 100G backbone with other participants. Initially, Indiana University and Purdue University will have 100G connections.

“This is an extremely important milestone that advances the strategic alignment between the high performance computing and Internet2 communities,” said Indiana University President Michael A. McRobbie. “Having the massive computational capabilities of XSEDE coupled directly to the most advanced network capabilities with the broadest reach among the U.S. and global research communities is of critical importance to the success of XSEDE.”

“XSEDE will now be able to provide the fastest research and education network in the nation to more than 8,000 scientists who regularly use some of the fastest supercomputers in the world,” said Internet2 President and CEO H. David Lambert. “Our new network allows researchers to conduct the latest scientific research in ways that they have not been able to do previously. I am excited to hear about future research impacting society from scientists using XSEDE resources and our new network.”

With the migration to Internet2, new sites can be easily added to the XSEDE backbone through connections to Internet2 Advanced Layer 2 Service nodes. In addition, the upgrade also eliminates the potential bottleneck between Denver and Chicago associated with the prior XSEDE backbone architecture. The new network infrastructure is expected to enable the development of new applications, such as an XSEDE-wide File System (XWFS), which will allow the increasingly very large files required by researchers to be moved rapidly between XSEDE sites.

About XSEDE
XSEDE, the most advanced, powerful and robust collection of integrated digital resources and services in the world, is a single virtual system that scientists can use to interactively share computing resources, data and expertise. The five-year, $121 million project is supported by the National Science Foundation. More about XSEDE: https://xsede.org.

About Internet2
Internet2® is a member-owned advanced technology community founded by the nation’s leading higher education institutions in 1996. Internet2 provides a collaborative environment for U.S. research and education organizations to solve common technology challenges, and to develop innovative solutions in support of their educational, research, and community service missions.

Internet2 consists of more than 220 U.S. universities, 60 leading corporations, 70 government agencies, 38 regional and state education networks and more than 100 national research and education networking partners representing more than 50 countries. Internet2 offices are located in Ann Arbor, Mich.; Emeryville, Calif.; and Washington, DC. For more information, visit www.internet2.edu.

RFP for Biomolecular Simulation Time on Anton due by Tuesday, May 28

Posted on by

The National Center for Multiscale Modeling of Biological Systems (MMBioS), and the National Resource for Biomedical Supercomputing (NRBSC) at the Pittsburgh Supercomputing Center (PSC) are soliciting proposals for computer time on Anton, a special-purpose supercomputer for molecular dynamics (MD) simulation designed by D. E. Shaw Research (DESRES).

Anton is nearly two orders of magnitude faster for doing molecular dynamics simulations than the previous state of the art supercomputers, enabling researchers to explore very long time scale dynamics. Anton can handle systems containing up to 150,000 atoms.

A 512-node Anton machine is currently available and in production use at the PSC. The machine has been made available without cost by DESRES for non-commercial research use by universities and other not-for-profit institutions, and is hosted by NRBSC. Operational funding is provided by NIGMS through MMBioS, a joint effort between the University of Pittsburgh, PSC, Carnegie Mellon University, and the Salk Institute.

To qualify for an allocation on Anton, the principal investigator (PI) must be a faculty or staff member at a U.S. academic or non-profit research institution and must have the appropriate level of authority and responsibility to direct the project supported by the allocation.
A graduate student or postdoctoral researcher may not be a PI, but a qualified advisor may apply for an allocation on her or his behalf. Furthermore, each investigator can serve as a PI for only a single application for computer time on Anton. PIs who previously had an allocation on Anton at NRBSC are strongly encouraged to apply but are required to provide a detailed (2-3 page) progress report of their past work on the Anton machine at NRBSC. Repeat applications unaccompanied by a progress report will not be considered.

The deadline for applications is 11:59 p.m., EDT, Tuesday, May 28, 2013.

There will be a workshop at PSC in the fall for people who are awarded allocations.

For complete details, including proposal requirements and submission instructions, see the NRBSC web site at http://www.nrbsc.org/anton_rfp/.