Berkeley Lab to have Strong Presence at SC11 Conference
November 10, 2011
Once again, scientists and engineers from Berkeley Lab are making significant contributions to the SC11 Technical Program, sharing their expertise and experience with thousands of attendees at the annual conference sponsored by the IEEE Computer Society and ACM SIGARCH. SC11 will be held Nov. 12-16 in Seattle.
SC11 Technical Program Sessions with Berkeley Lab Contributors
The SC11 Technical Papers program received 352 high quality submissions covering a variety of advanced research topics in HPC spanning six areas--Applications, Architecture/Networks, Clouds and Grids, Performance, Storage, and Systems Software. After an extremely rigorous peer review process, 74 papers were accepted for presentation. Berkeley Lab researchers contributed to 11 of those papers, including one nominated for Best Paper Award.
Tuesday, Nov. 15
“Multithreaded Global Address Space Communication Techniques for Gyrokinetic Fusion Applications on Ultra-Scale Platforms,” Robert Preissl, LBNL; Nathan Wichmann, Bill Long, Cray Inc.; John Shalf, LBNL; Stephane Ethier, PPPL; Alice Koniges, LBNL. (1:30 p.m.) BEST PAPER AWARD NOMINEE
“Gyrokinetic Toroidal Simulations on Leading Multi- and Manycore HPC Systems,” Kamesh Madduri, Khaled Ibrahim, Samuel Williams; Eun-Jin Im, Kookmin University; Stephane Ethier, PPPL; John Shalf and Leonid Oliker, LBNL. (3:30 p.m.)
“Unitary Qubit Lattice Simulations of Multiscale Phenomena in Quantum Turbulence,” George Vahala, College of William and Mary; Min Soe, Rogers State University; Bo Zhang, College of William and Mary; Jeffrey Yepez, Air Force Research Laboratory; Linda Vahala, Old Dominion University; Jonathan Carter, LBNL; and Sean Ziegeler, HPTi. (4 p.m.)
“Parallel Index and Query for Large Scale Data Analysis,” Jerry Chou, Kesheng Wu, Mark Howison, Prabhat, Oliver Ruebel, Brian Austin, E. Wes Bethel, Ji Qiang, Robert Ryne and Arie Shoshani, LBNL. (4 p.m.)
Wednesday, Nov. 16
“Multi-Science Applications with Single Codebase - GAMER - for Massively Parallel Architectures,” Hemant Shukla, LBNL; Hsi-Yu Schive, Tak-Pong Woo and Tzihong Chiueh, National Taiwan University. (11:30 a.m.)
“Optimized Pre-Copy Live Migration for Memory Intensive Applications,” Khaled Ibrahim, Costin Iancu, Steven Hofmeyr and Eric Roman, LBNL. (11:30 a.m.)
“Efficient Data Race Detection for Distributed Memory Parallel Programs,” Chang-Seo Park, University of California, Berkeley; Paul Hargrove, Costin Iancu, LBNL; Koushik Sen, UC Berkeley. (4 p.m.)
Thursday, Nov. 17
“Extracting Ultra-Scale Lattice Boltzmann Performance via Hierarchical and Distributed Auto-Tuning,” Samuel Williams, Leonid Oliker, Jonathan Carter and John Shalf, LBNL. (11 a.m.)
“Large Scale Plane Wave Pseudopotential Density Functional Theory Calculations on GPU Clusters,” Long Wang, Weile Jia, Xuebin Chi, Supercomputing Center of Computer Network Information Center, Chinese Academy of Sciences; Yue Wu and Weiguo Gao, MOE Key Laboratory of Computational Physical Sciences, School of Mathematical Sciences, Fudan University; Lin-Wang Wang, LBNL. (2:30 p.m.)
“Parallel Breadth-First Search on Distributed Memory Systems,” Aydin Buluc and Kamesh Madduri, LBNL. (2:30 p.m.)
“Hardware, Software Co-design for Energy Efficient Seismic Modeling,” Jens Krueger, Fraunhofer ITWM; David Donofrio, John Shalf, Samuel Williams, Leonid Oliker, LBNL; Marghoob Mohiyuddin, UC Berkeley; Franz-Josef Pfreundt, Fraunhofer ITWM. (4 p.m.)
A total of 68 proposed tutorials were submitted and 30 were accepted. will be presented at the conference. Of those, Berkeley Lab staff will present in three.
Linear Algebra Libraries for High-Performance Computing: Scientific Computing with Multicore and Accelerators (S08)
8:30 a.m. – 5 p.m. Sunday, Nov. 13
Presenters :Jack Dongarra, University of Tennessee, Knoxville; James Demmel, Berkeley Lab and UC Berkeley; Michael Heroux, Sandia National Laboratories; Jakub Kurzak, University of Tennessee.
Intro to PGAS (UPC and CAF) and Hybrid for Multicore Programming (M08)/
8:30 a.m. – 5 p.m. Monday, Nov. 14
Presenters: Alice Koniges and Katherine Yelick, Berkeley Lab/NERSC; Rolf Rabenseifner, High Performance Computing Center Stuttgart; Reinhold Bader, Leibniz Supercomputing Centre; David Eder, Lawrence Livermore National Laboratory.
Parallel I/O in Practice (M09)
8:30 a.m. – 5 p.m. Monday, Nov. 14
Presenters: Robert Latham and Robert Ross, Argonne National Laboratory; Brent Welch, Panasas; Katie Antypas, Lawrence Berkeley National Laboratory/NERSC
“Scaling to Petascale and Beyond: Performance Analysis and Optimization of Applications,” Presenters: Glenn Brook, Donald Frederick, Richard Gerber, Lawrence Berkeley National Laboratory/NERSC Jeff Larkin
A total of 176 posters were submitted, of which 70 were selected for presentation at the conference.
“I/O Workload Analysis with Server-side Data Collection”
Authors: Andrew C. Uselton and Daniela Ushizima, Berkeley Lab
Chaired by Harvey Wasserman, Berkeley Lab/NERSC, the SC11 panel sessions promise a variety of lively debates on active research topics and discussions centered on successful approaches to HPC problems.
“Energy Efficient Networks in Next Generation Data and Compute Centers”
10:30 a.m. – 12 p.m. Wednesday, Nov. 16
Moderator: Wael Diab, Broadcom
Panelists: Brent Draney, Berkeley Lab/NERSC; Keren Bergman, Columbia University; David Law, Hewlett-Packard; Kannan Raj, Oracle Labs
“Scientific Data on the Path to Exascale: Lessons, Insights and Predictions from 10+ years on the front lines”
3:30 p.m. – 5 p.m. Wednesday, Nov. 16
Moderator: Terence Critchlow, Pacific Northwest National Laboratory
Panelists: Rob Ross, Argonne National Laboratory; Nagiza Samatova, Oak Ridge National Laboratory; Lucy Nowell, DOE Office of Advanced Scientific Computing Research; Arie Shoshani, Berkeley Lab
“The View: HPC Edition”
8:30 – 10 a.m. Friday, Nov. 18
Moderator: Cherri M. Pancake, Oregon State University
Panelists: Debra Goldfarb, Microsoft Corp.; Alice Koniges, Berkeley Lab/NERSC; Candy Culhane, National Security Agency.
Workshops provide attendees with independently planned full-, half-, or multi-day sessions that complement the SC11Technical Program and extend its impact by providing greater depth of focus.
“Data Centers Have Gone Green (Or Haven't They?). When Will Networks Follow?”
9 a.m. – 5 p.m. Sunday, Nov. 13
Organizers: Thierry Klein and Dan Kilper, Alcatel-Lucent; Inder Monga, Berkeley Lab/ESnet
“Network-Aware Data Management”
9 a.m. – 5 p.m. Monday, Nov. 14
Organizers; Mehmet Balman, Surendra Byna and Arie Shoshani, Berkeley Lab
“Fourth Workshop on High Performance Computational Finance”
E. Wes Bethel, David Leinweber, Oliver Ruebel and Kesheng John Wu will present their paper on “Federal Market Information Technology in the Post Flash Crash Era: Roles for Supercomputing” as part of the workshop.
The Birds of a Feather (BOF) sessions are among the most interactive, popular, and well-attended sessions of the SC Conference Series. The BOF sessions provide a non-commercial, dynamic venue for conference attendees to openly discuss topics of focused mutual interest and currency within the HPC community with a strong emphasis on audience-driven discussion, professional networking and grass-roots participation.
“Setting Trends for Energy Efficiency”
12:15 – 1:15 p.m. Tuesday, Nov. 15
Primary Session Leader: Erich Strohmaier, Berkeley Lab
Session Co-leaders: Wu Feng, Virginia Tech; Nicolas Dube, Hewlett-Packard
“‘Hot’ for Warm-Water Cooling”
5:30 – 7 p.m. Tuesday, Nov. 15
Session Leaders: William Tschudi, Berkeley Lab
Session Co-leaders: David Martinez, Sandia National Laboratories; Farhad Banisadr, Los Alamos National Laboratory
5:30 – 7 p.m. Tuesday, Nov. 15
Session Leader: Erich Strohmaier, Berkeley Lab
12:15 – 1:15 p.m. Wednesday, Nov. 16
Primary Session Leader: Jon Dugan, Berkeley Lab/ESnet
Co-leader: Aaron Brown, Internet2
“Network Service Interface: Providing virtualized, dynamic services over the Wide Area Network”
5:30 – 7 p.m. Wednesday, Nov. 16
Primary Session Leader: Inder Monga, Berkeley Lab/ESnet
Co-leaders: Tomohiro Kudoh, National Institute of Advanced Industrial Science & Technology; Guy Roberts, DANTE
“Terabit Networks: Opportunity and Challenges in Extreme-Scale Science and Massive Data Movement”
12:15 – 1:15 p.m. Thursday, Nov. 17
Primary Session Leader: Thomas Ndousse, DOE Office of Advanced Scientific Computing Research
Co-Leader: Steve Cotter, Berkeley Lab/ESnet
BROADER ENGAGEMENT PROGRAM PRESENTATION
The goal of the Broader Engagement (BE) Program is to increase the participation of individuals who have been traditionally underrepresented in high performance computing (HPC). The program offers special activities to introduce, engage and support a diverse community in the conference and in HPC.
“HPC Challenges and Directions”
1:30 – 3 p.m. Monday, Nov. 14
Presenters: Horst Simon, Berkeley Lab; Dan Reed, Microsoft Research
SCINET RESEARCH SANDBOX
The SCinet Research Sandbox (SRS) will allow researchers with innovative network approaches to experimentally test their ideas in the unique environment of the SCinet network.
“End-to-End Virtualization – Campus, WAN and Data Center”
10:30 – 10:45 a.m. Thursday, Nov. 17
Presenters: Inder Monga, Eric Pouyoul and Brian Tierney, Berkeley Lab/ESnet; Ezra Kissel and Martin Swany, University of Delaware
STATE OF THE PRACTICE
(Chaired by David Paul, Berkeley Lab/NERSC)
An exciting new element for SC11, the State of the Practice provides a venue for the discussion and exchange of best practices involving provisioning, using and improving the critical systems and services in high performance computing, networking, and storage.
“Intrusion Detection at 100G”
2 – 2:30 p.m. Wednesday, Nov. 16
Authors: Scott Campbell and Jason Lee, Berkeley Lab/NERSC
“LOGJAM: A Scalable Unified Log File Archiver”
2:30 – 3 p.m. Thursday, Nov. 17
Author: Nicholas P. Cardo, Berkeley Lab/NERSC
SC11 Berkeley Lab Booth Activities
Booth 512 (Fourth Floor, Washington State Convention Center)
Featured Talks and Roundtable Discussions
Monday, Nov. 14
7-8 p.m.: A toast to ESnet’s 25 years of networking leadership and the Advanced Networking Initiative
Demonstration of 100 Gbps capability using astrophysics data
Tuesday, Nov. 15
10:30-11:15 a.m.: Featured talk on “Five Myths of Cloud Computing Debunked,” by Shane Canon, NERSC
11:15 a.m.-12:15 pm.: Roundtable discussion of the Magellan Cloud Computing Project led by Shane Canon and Lavanya Ramakrishnan, NERSC/CRD
1-2 p.m.: Roundtable discussion of “MyESnet Portal: Real Time Insight into Network’s Inner Workings” led by Jon Dugan, ESnet
2-3 p.m.: Roundtable discussion of” Berkeley Lab Checkpoint/Restart” led by Paul Hargrove, CRD Future Technologies Group
3-4 p.m.: Roundtable discussion of “ESnet’s Green Networking Initiative” led by Inder Monga, ESnet
Wednesday, Nov. 16
10:30-11:15 a.m.: Featured talk on “ESnet's Advanced Networking Initiative - Accelerating Science to 100G” by Steve Cotter, ESnet
11:15 a.m.-Noon: Featured talk on “Supernova Discovery in the Era of Data-Intensive Science” by Rollin Thomas, CRD Computational Cosmology Center
1-1:45 p.m.: Featured talk on “CoDesign for Exascale (CoDEx): Architectural Simulation and Modeling for Exascale Platform Development” by John Shalf, CRD/NERSC
2-3 p.m.: Roundtable discussion on “A Network Runtime Library for PGAS Languages (GASNet)” led by Paul Hargrove, CRD Future Technologies Group
3-4 p.m.: Roundtable discussion of “Warewulf Cluster Management” s led by Greg Kurtzer and Gary Jung, IT Division
4-5 p.m.: Roundtable discussion of “Careers at Berkeley Lab” led by Jeff Todd, CS Recruiter
Thursday, Nov. 17
11 a.m.-noon: Roundtable discussion of” Roles for Supercomputing in the Financial Markets” led by David Leinweber, CRD Center for Innovative Financial Technology
Ongoing Booth Demonstrations
Science at 100G led by Prabhat. This demo is intended to demonstrate the power of 100G networking when combined the power of NERSC supercomputing and Magellen cloud resources. he demo will leverage a 100G connection to the exhibit floor to show a high impact visualization of cosmology data, data pre-computed on Hopper, moved to Magellan IO nodes.
Climate Application on 100Gbps Network Demo led by Alex Sim and Mehmet Balman. Using the MyESnet portal and the ANI 100G prototype network, this demo will showcase the ability to transfer climate data at unprecedented speeds.
Drag and Drop Supercomputing led by Annette Greiner and Iwona Sakrejda (NERSC). With the NERSC Online VASP Application (NOVA), we have built a graphical interface with high usability that lets NERSC users set up, run, and monitor quantum mechanical calculations using VASP. In this hands-on, walk-up demo, visitors can try running and examining DFT simulations through a web interface.
Interactive CoDEx demo led by David Donofrio. Test your processor design skills as you change processor architecture parameters and then see the impact of your design choices on power and performance on a few scientific kernels.
SCinet Research Sandbox/OpenFlow/OSCARS demo led by Eric Pouyoul and Brian Tierney. This demonstration will showcase the ability to bring together OSCARS and OpenFlow to create a seamless, automated, end-to-end virtualized path with guaranteed bandwidth capabilities to enable high-speed data movement between Campus and Data Centers over the WAN infrastructure.
About Computing Sciences at Berkeley Lab
The Computing Sciences Area at Lawrence Berkeley National Laboratory provides the computing and networking resources and expertise critical to advancing Department of Energy Office of Science research missions: developing new energy sources, improving energy efficiency, developing new materials, and increasing our understanding of ourselves, our world, and our universe.
Founded in 1931 on the belief that the biggest scientific challenges are best addressed by teams, Lawrence Berkeley National Laboratory and its scientists have been recognized with 13 Nobel Prizes. Today, Berkeley Lab researchers develop sustainable energy and environmental solutions, create useful new materials, advance the frontiers of computing, and probe the mysteries of life, matter, and the universe. Scientists from around the world rely on the Lab’s facilities for their own discovery science. Berkeley Lab is a multiprogram national laboratory, managed by the University of California for the U.S. Department of Energy’s Office of Science.
DOE’s Office of Science is the single largest supporter of basic research in the physical sciences in the United States, and is working to address some of the most pressing challenges of our time. For more information, please visit energy.gov/science.