Computing and Computational Sciences Directorate
SuperComputing 2011
Printed Materials | Videos | Committees | Technical Papers | Tutorials | Broader Engagement |Panels | Posters | Masterworks | BOFs | Gordon Bell Finalists | Awards Won | Workshops | Related Activities | Booth Speakers | Booth Model Images | Kiosk Presentations
SYSTEMS AND FACILITIES
- GAEA Overview
- Jaguar Overview Pt. 1
- Jaguar Overview Pt. 2
- Kraken Overview
- NICS Overview
- OLCF: The Home of Game-Changing Science
- Spider Lustre File System a Trailblazer at OLCF
SCIENCE
- Climate Tackles Clouds
- Earth Quake Simulation Rocks Southern California
- OLCF Supercomputing Experts
- Resolving Global Climate Change
- Supercomputer Sheds Light on HIV’s Behavior
- Supercomputers Simulate the Molecular Machines that Replicate and Repair DNA
- Magnitude-8 Earthquake Simulation
- Lithium Air Battery Technology
- Nanoscale Simulations
- Shock Compression Technology for Carbon Sequestration
- Turbomachines
- Numerical Model of Class A Tractor Trailer
ORNL/UT participation in SC2011 Committees
- Conference Vice-Chair: Becky Verastegui, Oak Ridge National Laboratory
- Technical Program Executive Director: James H. Rogers, Oak Ridge National Laboratory
- Past Chair:
- Ricky A. Kendall, Oak Ridge National Laboratory
- System Software Committee Member: Arthur Maccabe, Oak Ridge National Laboratory
- Applications Committee Member:
- Ricky A. Kendall, Oak Ridge National Laboratory
- Scott Klasky, Oak Ridge National Laboratory
- Douglas Kothe, Oak Ridge National Laboratory
- Architecture and Networks Committee Member: Rich Graham, Oak Ridge National Laboratory
- Storage Committee Member: Xiaosong Ma, North Carolina State University / Oak Ridge National Laboratory
- Birds of a Feather Chair: Jeffery A. Kuehn, Oak Ridge National Laboratory
- Posters Co-Chair: Philip C. Roth, Oak Ridge National Laboratory
- Architectures/Networks Area Chair: Arthur Maccabe, Oak Ridge National Laboratory
- Architectures/Networks Committee Member: Collin B. McCurdy, Oak Ridge National Laboratory
- Tutorials Committee Member:
- John W. Cobb, Oak Ridge National Laboratory
- Bradley Settlemyer, Oak Ridge National Laboratory
- Jeffrey Vetter, Oak Ridge National Laboratory and Georgia Tech
- Patrick Haven Worley, Oak Ridge National Laboratory
- Student Cluster Challenge Chair: Douglas Fuller, Oak Ridge National Laboratory
- Broader Engagement Committee Member:
- Rebecca Hartman-Baker, Oak Ridge National Laboratory
- Debbie McCoy, Oak Ridge National Laboratory
- BE Scavenger Hunt Chair: Rebecca Hartman-Baker, Oak Ridge National Laboratory
- Student Cluster Challenge Chair: Douglas Fuller, Oak Ridge National Laboratory
- Student Cluster Competition Committee Member: Dustin Leverman, Oak Ridge National Laboratory
- Student Cluster Competition Advisory Board Member: Ricky A. Kendall, Oak Ridge National Laboratory
- SCinet Chair Emeritus: James H. Rogers, Oak Ridge National Laboratory
- Architecture Co-Chair: Charles D. Fisher, Oak Ridge National Laboratory
- Commodity Co-Chair: Rex Duncan, Oak Ridge National Laboratory
- Fiber Team Member:
- DA Fye, Oak Ridge National Laboratory
- Zachary Giles, Oak Ridge National Laboratory
- Executive Committee:
- James H. Rogers, Oak Ridge National Laboratory
- Becky Verastegui, Oak Ridge National Laboratory
ORNL/UT participation in SC2011 Technical Papers
Tuesday
Optimizing Symmetric Dense Matrix-Vector Multiplication on GPUs
SESSION: Dense Linear Algebra
EVENT TYPE: Paper
TIME: 10:30AM - 11:00AM
AUTHOR(S):Rajib Nath, Stanimire Tomov, Tingxing Dong, Jack Dongarra
ROOM:TCC 305
ABSTRACT:
GPUs are an excellent accelerator for data-parallel applications with regular data access patterns. It is challenging, however, to optimize computations with irregular data access patterns on GPUs. One such computation is the Symmetric Matrix Vector product (SYMV) for dense linear algebra. Optimizing the SYMV kernel is important be- cause it forms the basis of fundamental algorithms such as linear solvers and eigenvalue solvers on symmetric matrices. In this work, we present a new algorithm for optimizing the SYMV kernel on GPUs. Our optimized SYMV in single precision brings up to 7X speed up compared to the (latest) CUBLAS 3.2 NVIDIA library on the GTX 280 GPU. Our SYMV kernel tuned for Fermi C2050 is 4.5X faster than CUBLAS 3.2 in single precision.
Chair/Author Details:
- Rajib Nath - University of California, San Diego
- Stanimire Tomov - University of Tennessee, Knoxville
- Tingxing Dong - University of Tennessee, Knoxville
- Jack Dongarra - University of Tennessee, Knoxville
Simplified Parallel Domain Traversal
SESSION: Domain Specific Languages
EVENT TYPE: Paper, Best Student Paper (BSP) Finalist
TIME: 11:00AM - 11:30AM
AUTHOR(S):Wesley Kendall, Jingyuan Wang, Melissa Allen, Tom Peterka, Jian Huang, David Erickson
ROOM:TCC 304
ABSTRACT:
Many data-intensive scientific analysis techniques require global domain traversal, which over the years has been a bottleneck for efficient parallelization across distributed-memory architectures. Inspired by MapReduce and other simplified parallel programming approaches, we have designed DStep, a flexible system that greatly simplifies efficient parallelization of domain traversal techniques at scale. In order to deliver both simplicity to users as well as scalability on HPC platforms, we introduce a novel two-tiered communication architecture for managing and exploiting asynchronous communication loads. We also integrate our design with advanced parallel I/O techniques that operate directly on native simulation output. We demonstrate DStep by performing teleconnection analysis across ensemble runs of terascale atmospheric CO2 and climate data, and we show scalability results on up to 65,536 IBM BlueGene/P cores.
Chair/Author Details:
- Wesley Kendall - University of Tennessee, Knoxville
- Jingyuan Wang - University of Tennessee, Knoxville
- Melissa Allen - University of Tennessee, Knoxville
- Tom Peterka - Argonne National Laboratory
- Jian Huang - University of Tennessee, Knoxville
- David Erickson - Oak Ridge National Laboratory
Parallel Reduction to Condensed Forms for Symmetric Eigenvalue Problems using Aggregated Fine-Grained and Memory-Aware Kernels
SESSION: Dense Linear Algebra
EVENT TYPE: Paper
TIME: 11:30AM - 12:00PM
AUTHOR(S):Azzam Haidar, Hatem Ltaief, Jack Dongarra
ROOM:TCC 305
ABSTRACT:
This paper introduces a novel implementation in reducing a symmetric dense matrix to tridiagonal form, which is the preprocessing step toward solving symmetric eigenvalue problems. Based on tile algorithms, the reduction follows a two-stage approach, where the tile matrix is first reduced to symmetric band form prior to the final condensed structure. The challenging trade-off between algorithmic performance and task granularity has been tackled through a grouping technique, which consists in aggregating fine-grained and memory-aware computational tasks during both stages, while sustaining the application overall high performance. A dynamic runtime environment system schedules then the different tasks in an out-of-order fashion. The performance for the tridiagonal reduction reported in this paper are unprecedented. Our implementation results in an up to 50-fold improvement (125 Gflop/s) compared to the equivalent routine from LAPACK and Intel MKL on an eight socket hexa-core AMD Opteron multicore shared-memory system with a matrix size of 24000x24000.
Chair/Author Details:
- Azzam Haidar - University of Tennessee, Knoxville
- Hatem Ltaief - KAUST Supercomputing Laboratory
- Jack Dongarra - University of Tennessee, Knoxville
An Image Compositing Solution at Scale
SESSION: Applications
EVENT TYPE: Paper
TIME: 4:30PM - 5:00PM
AUTHOR(S):Kenneth Moreland, Wesley Kendall, Tom Peterka, Jian Huang
ROOM:TCC 304
ABSTRACT:
The only proven method for performing distributed-memory parallel rendering at large scales, tens of thousands of nodes, is a class of algorithms called sort last, which combines a collection of images generated independently on each node into a single blended image. Over the years numerous image compositing algorithms have been proposed as well as several enhancements and rendering modes to these core algorithms. However, the testing of these algorithms has been with arbitrary, if any, enhancements. This paper takes a leading production-quality image-compositing framework, IceT, and uses it as a testing framework for the leading image compositing algorithms of today. As we scale IceT to ever increasing job sizes, we consider the image compositing systems holistically, incorporate numerous optimizations, and discover several improvements to the process never considered before. We conclude by demonstrating our solution on 64K cores of the Intrepid BlueGene/P at Argonne National Laboratories.
Chair/Author Details:
- Kenneth Moreland - Sandia National Laboratories
- Wesley Kendall - University of Tennessee, Knoxville
- Tom Peterka - Argonne National Laboratory
- Jian Huang - University of Tennessee, Knoxville
ISABELA-QA: Query-driven Data Analytics over ISABELA-compressed Extreme-Scale Scientific Data
SESSION: Querying Large Scale Data
EVENT TYPE: Paper
TIME: 4:30PM - 5:00PM
AUTHOR(S):Sriram Lakshminarasimhan, Jonathan Jenkins, Robert Latham, Robert Ross, Nagiza F. Samatova, Isha Arkatkar, Zhenhuan Gong, Hemanth Kolla, Jackie Chen, Seung-Hoe Ku, C.S. Chang, Stephane Ethier, Scott Klasky
ROOM:TCC 305
ABSTRACT:
We present a query processing engine for scientific data based on ISABELA, a partitioned B-spline-lossy-compression scheme. We optimize spatial region and variable queries on variable and temporal constraints by performing temporal-delimited binning on the range of the variable values, ensuring near-uniform distribution of compressed data across bins. We demonstrate the high, user-controlled accuracy of reconstructed data through several analytic scenarios, and the competitive performance of variable/temporal-constrained query processing, while incurring both a smaller memory and storage footprint compared to both raw data and popular scientific database systems. Finally, we discuss a number of HPC optimizations, such as parallel I/O and multi-node/multi-core query processing parallelized by temporal constraints, that allow extreme scale scalability.
Chair/Author Details:
- Sriram Lakshminarasimhan - North Carolina State University
- Jonathan Jenkins - North Carolina State University
- Robert Latham - Argonne National Laboratory
- Robert Ross - Argonne National Laboratory
- Nagiza F. Samatova - Oak Ridge National Laboratory
- Isha Arkatkar - North Carolina State University
- Zhenhuan Gong - North Carolina State University
- Hemanth Kolla - Sandia National Laboratories
- Jackie Chen - Sandia National Laboratories
- Seung-Hoe Ku - New York University
- C.S. Chang - New York University
- Stephane Ethier - Princeton Plasma Physics Laboratory
- Scott Klasky - Oak Ridge National Laboratory
Thursday
Performance of the Community Earth System Model
SESSION: Application Performance
EVENT TYPE: Paper
TIME: 10:30AM - 11:00AM
AUTHOR(S):Patrick H. Worley, Anthony P. Craig, John M. Dennis, Arthur A. Mirin, Mark A. Taylor, Mariana Vertenstein
ROOM:TCC 303
ABSTRACT:
The Community Earth System Model (CESM), released in June 2010, incorporates new physical process and new numerical algorithm options, significantly enhancing simulation capabilities over its predecessor, the June 2004 release of the Community Climate System Model. CESM also includes enhanced performance tuning options and performance portability capabilities. This paper describes performance and performance scaling on both the Cray XT5 and the IBM BG/P for four representative production simulations, varying both problem size and enabled physical processes. The paper also describes preliminary performance results for high resolution simulations using over 200,000 processor cores, indicating the promise of ongoing work in numerical algorithms and where further work is required.
Chair/Author Details:
- Patrick H. Worley - Oak Ridge National Laboratory
- Anthony P. Craig - National Center for Atmospheric Research
- John M. Dennis - National Center for Atmospheric Research
- Arthur A. Mirin - Lawrence Livermore National Laboratory
- Mark A. Taylor - Sandia National Laboratories
- Mariana Vertenstein - National Center for Atmospheric Research
Scalable Implementations of Accurate Excited-state Coupled Cluster Theories: Application of High-level Methods to Porphyrin-based Systems
SESSION: Applications
EVENT TYPE: Paper
TIME: 3:30PM - 4:00PM
AUTHOR(S):Karol Kowalski, Sriram Krishnamoorthy, Ryan Olson, Vinod Tipparaju, Eduardo Apra
ROOM:TCC 304
ABSTRACT:
The development of reliable tools for excited-state simulations is important for understanding complex processes in the broad class of light harvesting systems and optoelectronic devices. Over last years we have been developing equation of motion coupled cluster(EOMCC) methods capable of tackling these problems. In this paper we discuss parallel performance of EOMCC codes which provide accurate description of excited-state correlation effects. Two aspects are discussed in detail:(1) a new algorithm for the iterative EOMCC methods based on improved parallel task scheduling algorithms, and (2) parallel algorithms for the non-iterative methods describing the effect of triply excited configurations. We demonstrate that the most computationally intensive non-iterative part can take advantage of 210,000 cores of the Cray XT5 system at the Oak Ridge Leadership Computing Facility(OLCF), achieving over 80% parallel efficiency. In particular, we demonstrate importance of computationally demanding non-iterative many-body methods in matching experimental level of accuracy for several porphyrin-based systems.
Chair/Author Details:
- Karol Kowalski - Pacific Northwest National Laboratory
- Sriram Krishnamoorthy - Pacific Northwest National Laboratory
- Ryan Olson - Cray Inc.
- Vinod Tipparaju - Oak Ridge National Laboratory
- Eduardo Apra - Oak Ridge National Laboratory
ORNL/UT participation in SC2011 Tutorials
Sunday November 13
S08: Linear Algebra Libraries for High-Performance Computing: Scientific Computing with Multicore and Accelerators
SESSION: S08: Linear Algebra Libraries for High-Performance Computing: Scientific Computing with Multicore and Accelerators
EVENT TYPE: Tutorial
TIME: 8:30AM - 5:00PM
Presenter(s):Jack Dongarra, James Demmel, Michael Heroux, Jakub Kurzak
ROOM:
ABSTRACT:
Today, a desktops with a multicore processor and a GPU accelerator can already provide a TeraFlop/s of performance, while the performance of the high-end systems, based on multicores and accelerators, is already measured in PetaFlop/s. This tremendous computational power can only be fully utilized with the appropriate software infrastructure, both at the low end (desktop, server) and at the high end (supercomputer installation). Most often a major part of the computational effort in scientific and engineering computing goes in solving linear algebra subproblems. This tutorial surveys the state-of-the-art numerical libraries for solving problems in linear algebra, both dense and sparse. We highlight recent algorithms that minimize communication, i.e. moving data, which is much more expensive than arithmetic. The following software packages are presented: PLASMA & MAGMA for solving dense systems using multicores and accelerators, Trilinos for solving sparse systems using multicore, accelerators and distributed memory systems.
Chair/Presenter Details:
- Jack Dongarra - University of Tennessee, Knoxville
- James Demmel - University of California, Berkeley
- Michael Heroux - Sandia National Laboratories
- Jakub Kurzak - University of Tennessee
S10: Scaling to Petascale and Beyond: Performance Analysis and Optimization of Applications
SESSION: S10: Scaling to Petascale and Beyond: Performance Analysis and Optimization of Applications
EVENT TYPE: Tutorial
TIME: 8:30AM - 5:00PM
Presenter(s):Glenn Brook, Donald Frederick, Richard Gerber, Jeff Larkin
ROOM:
ABSTRACT:
Current supercomputer systems have architectures with many thousands of cores. Application developers are faced with the task of designing for parallel scalability at ever higher levels. The effective use of such platforms, requires application developers to focus on achieving parallel scalability through application design and optimization driven by performance analysis. This tutorial provides developers with the background in system features, program design and performance analysis to aid in producing efficient programs on modern supercomputers.
Chair/Presenter Details:
- Glenn Brook - University of Tennessee, Knoxville
- Donald Frederick - Lawrence Livermore National Laboratory
- Richard Gerber - Lawrence Berkeley National Laboratory
- Jeff Larkin - Cray Inc.
S14: In-Situ Visualization with ParaView
SESSION: S14: In-Situ Visualization with ParaView
EVENT TYPE: Tutorial
TIME: 1:30PM - 5:00PM
Presenter(s):Andrew C. Bauer, Nathan Fabian, Norbert Podhorszki, Pat Marion
ROOM:
ABSTRACT:
In-situ visualization is a term for running a solver in tandem with visualization. By coupling these together we can utilize the high performance computing for post-processing while circumventing the bottlenecks associated with storing and retrieving data in disk storage. ParaView is a powerful open-source turnkey application for analyzing and visualizing large data sets in parallel. We demonstrate two methods for integrating ParaView for in-situ visualization. The first method is through ParaView’s coprocessing library. It simplifies the integration with simulation codes by providing a programmatic interface to ParaView. Attendees will learn the structure of the coprocessing API and how to bind it to C, C++, Fortran, and Python. The second method uses the ADIOS I/O framework. By using the ADIOS API for I/O, simulation data can be written to disk or to a stage area hosted by another parallel application. Data analysis applications can fetch data directly from stage avoiding disk I/O. Attendees will learn the ADIOS API and how to use a staging method, which is integrated with ParaView. They will also get directions for writing ParaView plugins to read in their particular output. Attendees will also receive instructions on customizing the ParaView build on Leadership Computing Facilities architectures.
Chair/Presenter Details:
- Andrew C. Bauer - Kitware, Inc.
- Nathan Fabian - Sandia National Laboratories
- Norbert Podhorszki - Oak Ridge National Laboratory
- Pat Marion - Kitware, Inc.
Monday November 14
M12: Scalable Heterogeneous Computing on GPU Clusters
SESSION: M12: Scalable Heterogeneous Computing on GPU Clusters
EVENT TYPE: Tutorial
TIME: 8:30AM - 5:00PM
Presenter(s):Jeffrey Vetter, Kyle Spafford, Philip Roth, Allen Malony
ROOM:
ABSTRACT:
This tutorial, suitable for attendees with an intermediate-level in Parallel programming in MPI, and GPU programming in CUDA or OpenCL, will provide a comprehensive overview on the optimization techniques to port, analyze, and accelerate applications on scalable heterogeneous computing systems. We will focus on methods, tools, and techniques to migrate existing applications to large scale GPU clusters using MPI and OpenCL/CUDA. First, we will review our methodology for successfully identifying and selecting portions of applications to accelerate with a GPU, motivated with several application case studies. Second, we will present an overview of several performance and correctness tools, which provide performance measurement, profiling, and tracing information about applications running on these systems. Third, we will present a set of best practices for optimizing these applications: GPU and NUDA optimization techniques, optimizing interactions between MPI and GPU programming models. A hands-on session will be conducted on the NSF Keeneland Initial Delivery System, after each part to give participants the opportunity to investigate techniques and performance optimizations on such a system. Existing tutorial codes and benchmark suites will be provided to facilitate individual discovery. Additionally, participants may bring and work on their own applications.
Chair/Presenter Details:
- Jeffrey Vetter - Oak Ridge National Laboratory and Georgia Tech
- Kyle Spafford - Oak Ridge National Laboratory
- Philip Roth - Oak Ridge National Laboratory
- Allen Malony - University of Oregon
M13: Big Data Means Your Metadata Must Work
SESSION: M13: Big Data Means Your Metadata Must Work
EVENT TYPE: Tutorial
TIME: 1:30PM - 5:00PM
Presenter(s):Scott Jensen, Beth Plale, Rebecca Koskela, Amber Budden, John Cobb
ROOM:
ABSTRACT:
Data intensive computing means a lot of valuable data coming from parallel and graph computations and remote instruments. However, a recent Science article estimated that only 1% of ecological data is accessible after the research has been published. This tutorial addresses how metadata is critical to the use and reuse of scientific data. It discusses techniques for metadata capture, communication, and use in data discovery using four tools: XMC Cat, Metacat, Morpho, and Mercury. In this tutorial, participants will be taken through (1) installing and configuring a metadata catalog, (2) capturing domain-specific metadata both programmatically and using web-based user interfaces, (3) searching for metadata, and (4) configuring their search to return the right results. It will use examples from geo- and environmental sciences. At the conclusion of the tutorial, participants will have deeper familiarity with metadata tools currently available and will be on their way to making big data shareable!
Chair/Presenter Details:
- Scott Jensen - Indiana University
- Beth Plale - Indiana University
- Rebecca Koskela - University of New Mexico
- Amber Budden - DataONE
- John Cobb - Oak Ridge National Laboratory
ORNL/UT participation in SC2011 Broader Engagement
ORNL/UT participation in SC2011 Panels
Wednesday
The Path Forward on Programming Models, is there one?
SESSION: The Path Forward on Programming Models, is there one?
EVENT TYPE: Panel
TIME: 10:30AM - 12:00PM
SESSION CHAIR: Ricky Kendall
Panelists:Ricky Kendall, Rusty Lusk, John Levesque, Barbara Chapman, Michael Wolfe, Robert Harrison
ROOM:TCC 101
ABSTRACT:
We are at an architectural sea change in high performance computing. The Road Runner system over 2 years ago became the fastest computer in the world, the first hybrid system at the top of the TOP 500 list. The current leader is also a hybrid system, the Chinese Tianhe-1A system at the National Supercomputer Center in Tianjin at 2.57 Petaflops. The community as a whole will have to learn how to program and adapt science applications to these hybrid systems. This will need to be done in such a way that the overall modifications outlive these current machines and are easily adaptable to future architectures including the Exascale and beyond. What is the right programming model to use in our applications to not only enable this transformation in architecture but to facilitate portable performance? This panel will demonstrate options and opinions and open a dialog on this very important issue.
Moderator/Panelist Details:
- Ricky Kendall (Chair) - ORNL (testing pc member)
- Ricky Kendall (Moderator) - Oak Ridge National Laboratory
- Rusty Lusk - Argonne National Laboratory
- John Levesque - Cray Inc.
- Barbara Chapman - University of Houston
- Michael Wolfe - Portland Group
- Robert Harrison - Oak Ridge National Laboratory
Scientific Data on the Path to Exascale: Lessons, Insights and Predictions from 10+ years on the front lines
EVENT TYPE: Panel
TIME: 3:30PM - 5:00PM
SESSION CHAIR: Terence Critchlow
Panelists:Terence Critchlow, Rob Ross, Nagiza Samatova, Lucy Nowell, Arie Shoshani
ROOM:TCC 101
ABSTRACT:
This panel will present an overview of how data-centric research has made a difference to scientific applications within the SciDAC program, what advances have enabled data management and analysis to keep up with the increasing power of the supercomputers driving the simulations, and where the discipline is headed as we continue to move towards exascale computing. During their 20 minute presentations, the panelists will be encouraged to take strong positions while answering the following questions regarding data-centric research: - Give at least 2 examples of where data intensive research has enabled new scientific discoveries - What advancement has been the most disruptive to science over the past decade - What advancement has had the most impact on science over the past decade - What advancement has been most over-hyped within the past decade - What has not worked as well as you had expected - What advancement do you see as being the most important in the coming 5 years
Moderator/Panelist Details:
- Terence Critchlow (Chair) - PNNL
- Terence Critchlow (Moderator) - PNNL
- Rob Ross - Argonne National Laboratory
- Nagiza Samatova - Oak Ridge National Laboratory
- Lucy Nowell - DOE Office of Advanced Scientific Computing Research
- Arie Shoshani - Lawrence Berkeley National Laboratory
ORNL/UT participation in SC2011 Posters
Tuesday
New Features of the PAPI Hardware Counter Library
SESSION: Research Poster Reception
EVENT TYPE: ACM Student Research Competition Poster, Poster, Electronic Poster
TIME: 5:15PM - 7:00PM
SESSION CHAIR: Bernd Mohr
AUTHOR(S):Shirley Moore, Dan Terpstra, Vincent Weaver, Heike Jagode, James Ralph, Jack Dongarra
ROOM:WSCC North Galleria 2nd/3rd Floors
ABSTRACT:
The PAPI library has evolved from a cross-platform interface for accessing processor hardware performance counters to a component-based library for simultaneously accessing hardware monitoring information from various components of a computer system, including processors, memory controllers, network switches and interface cards, I/O subsystem, temperature sensors and power meters, and GPU counters. A GPU component is discussed. A new feature called user-defined events adds a layer of abstraction that allows users to define new metrics by combining previously defined events and machine constants and to share those metrics with other users. One current effort is the development of a PAPI interface for virtual machines, called PAPI-V, which will allow users to access processor and component hardware performance information from applications running within virtual machines. PAPI continues to be widely used by application developers and by higher level performance analysis tools such as TAU, PerfSuite, Scalasca, IPM, HPCtoolkit, Vampir, and CrayPat.
Chair/Author Details:
- Bernd Mohr (Chair) - Juelich Supercomputing Centre
- Shirley Moore - University of Tennessee, Knoxville
- Dan Terpstra - University of Tennessee, Knoxville
- Vincent Weaver - University of Tennessee, Knoxville
- Heike Jagode - University of Tennessee, Knoxville
- James Ralph - University of Tennessee, Knoxville
- Jack Dongarra - University of Tennessee, Knoxville
Scalable Infrastructure to Support Supercomputer Resiliency-Aware Applications and Load Balancing
SESSION: Research Poster Reception
EVENT TYPE: ACM Student Research Competition Poster, Poster, Electronic Poster
TIME: 5:15PM - 7:00PM
SESSION CHAIR: Bernd Mohr
AUTHOR(S):Yoav Tock, Benjamin Mandler, Jose Moreira, Terry Jones
ROOM:WSCC North Galleria 2nd/3rd Floors
ABSTRACT:
Current trends dictate increasing complexity and component counts on supercomputers and mainstream commercial systems alike. This trend exposes weaknesses in the underlying clustering infrastructure needed for continuous availability, maximizing utilization, and efficient administration of such systems. We propose a step to alleviate this problem by providing a highly scalable clustering infrastructure, based on peer-to-peer technologies, for supporting resiliency-aware applications as well as efficient monitoring and load balancing. Supported services include Membership, supporting resiliency-aware applications, publish-subscribe messaging, Convergecast supporting load balancing schemes, attributes dissemination for propagating slowly changing information, and DHT facilitating services discovery and integration. We employ a flexible two layer hierarchical topology, comprised of base zones federated by a management zone. Our design aspires to quickly identify membership changes in a scalable environment with minimal overall system disruption, thus enabling efficient Exascale size deployments. We present experimental evaluation taken from an IBM BlueGene/P, demonstrating scalability up to ~256K nodes.
Chair/Author Details:
- Bernd Mohr (Chair) - Juelich Supercomputing Centre
- Yoav Tock - IBM Research
- Benjamin Mandler - IBM Research
- Jose Moreira - IBM Research
- Terry Jones - Oak Ridge National Laboratory
DCA++ Study on Strong Correlated Electron System Case study: Nematic physics in Cuprate
SESSION: Research Poster Reception
EVENT TYPE: ACM Student Research Competition Poster, Poster, Electronic Poster
TIME: 5:15PM - 7:00PM
SESSION CHAIR: Bernd Mohr
AUTHOR(S):Shi-Quan Su, Thomas A. Maier, Michael S. Summers
ROOM:WSCC North Galleria 2nd/3rd Floors
ABSTRACT:
The DCA++ code implements a dynamic cluster quantum Monte Carlo (QMC) algorithm for the study of low-energy effective models of strongly correlated electron systems. It is written using a generic programming model and a Don't Repeat Yourself (DRY) philosophy and provides rich functionalities including flexible support for crystal structures, hybrid precision simulations as well as different QMC cluster solver algorithms. The DCA++ code provides a state-of-the-art tool to model and understand the physics of materials at the forefront of science. Here we describe an application to the cuprate high-temperature superconductors, which aims at unravelling the physics behind an electronic nematic phase, the spontaneous breaking of the rotational symmetry of the crystal, which has been observed in a number of experiments on these and related compounds.
Chair/Author Details:
- Bernd Mohr (Chair) - Juelich Supercomputing Centre
- Shi-Quan Su - Oak Ridge National Laboratory
- Thomas A. Maier - Oak Ridge National Laboratory
- Michael S. Summers - Oak Ridge National Laboratory
High-Level, One-Sided Programming Models on MPI: A Case Study with Global Arrays and NWChem
SESSION: Research Poster Reception
EVENT TYPE: ACM Student Research Competition Poster, Poster, Electronic Poster
TIME: 5:15PM - 7:00PM
SESSION CHAIR: Bernd Mohr
AUTHOR(S):James Dinan, Pavan Balaji, Jeff R. Hammond, Sriram Krishnamoorthy, Vinod Tipparaju
ROOM:WSCC North Galleria 2nd/3rd Floors
ABSTRACT:
Global Arrays (GA) is popular high-level parallel programming model that provides data and computation management facilities to the NWChem computational chemistry suite. GA's global-view data model is supported by the ARMCI partitioned global address space runtime system, which traditionally is implemented natively on each supported platform in order to provide the best performance. The industry standard Message Passing Interface (MPI) also provides one-sided functionality and is available on virtually every supercomputing system. We present the first high-performance, portable implementation of ARMCI using MPI one-sided communication. We interface the existing GA infrastructure with ARMCI-MPI and demonstrate that this approach reduces the amount of resources consumed by the runtime system, provides comparable performance, and enhances portability for applications like NWChem.
Chair/Author Details:
- Bernd Mohr (Chair) - Juelich Supercomputing Centre
- James Dinan - Argonne National Laboratory
- Pavan Balaji - Argonne National Laboratory
- Jeff R. Hammond - Argonne National Laboratory
- Sriram Krishnamoorthy - Pacific Northwest National Laboratory
- Vinod Tipparaju - Oak Ridge National Laboratory
A Novel Architecture for Biothreat Situation Awareness
SESSION: Research Poster Reception
EVENT TYPE: ACM Student Research Competition Poster, Poster, Electronic Poster
TIME: 5:15PM - 7:00PM
SESSION CHAIR: Bernd Mohr
AUTHOR(S):Thomas Brettin, Laura Pullum, Daniel Quest, Xiaohui Cui, Songhua Xu, Richard Stouder
ROOM:WSCC North Galleria 2nd/3rd Floors
ABSTRACT:
Data intensive computing has the potential to change biosurveillance. Biosurveillance requires the ability to ‘connect the dots’ and inform decision makers in near real-time. To do this effectively, computers need to handle information found on the web as diverse as DNA sequence and social media posts. A key challenge to process the web in near real-time is the rapid conversion of petabytes of raw information into a semantically rich and query-able knowledgebase that serves decision support needs. We present a novel architecture that combines: (1) parallel web crawlers guided by biological ontologies to extract relevant information, (2) a petascale cache to store, classify, index and cluster those portions of the web relevant to biosurveillance, and (3) a semantic store to integrate diverse sources of data, guide new searches and serve as a platform for data analysis and user applications. A poster and handouts will be available at the conference.
Chair/Author Details:
- Bernd Mohr (Chair) - Juelich Supercomputing Centre
- Thomas Brettin - Oak Ridge National Laboratory
- Laura Pullum - Oak Ridge National Laboratory
- Daniel Quest - Oak Ridge National Laboratory
- Xiaohui Cui - Oak Ridge National Laboratory
- Songhua Xu - Oak Ridge National Laboratory
- Richard Stouder - Oak Ridge National Laboratory
Detection and Correction of Silent Data Corruption for Large-Scale High-Performance Computing
SESSION: Research Poster Reception
EVENT TYPE: ACM Student Research Competition Poster, Poster, Electronic Poster
TIME: 5:15PM - 7:00PM
SESSION CHAIR: Bernd Mohr
AUTHOR(S):David Fiala, Frank Mueller, Christian Engelmann, Rolf Riesen, Kurt Ferreira
ROOM:WSCC North Galleria 2nd/3rd Floors
ABSTRACT:
Faults have become the norm rather than the exception for high-end computing on clusters with 10s/100s of thousands of cores. Exacerbating this situation, some of these faults will not be detected, manifesting themselves as silent errors that will corrupt memory while applications continue to operate and report incorrect results. This poster introduces RedMPI, an MPI library which resides in the MPI profiling layer. RedMPI is capable of both online detection and correction of soft errors that occur in MPI applications without requiring any modifications to the application source. By providing redundancy, RedMPI is capable of transparently detecting corrupt messages from MPI processes that become faulted during execution. Furthermore, with triple redundancy RedMPI additionally ``votes'' out MPI messages of a faulted process by replacing corrupted results with corrected results from unfaulted processes. We present an experimental evaluation of RedMPI on an assortment of applications to demonstrate the effectiveness of this approach.
Chair/Author Details:
- Bernd Mohr (Chair) - Juelich Supercomputing Centre
- David Fiala - North Carolina State University
- Frank Mueller - North Carolina State University
- Christian Engelmann - Oak Ridge National Laboratory
- Rolf Riesen - IBM
- Kurt Ferreira - Sandia National Laboratories
A Tunable, Software-based DRAM Error Detection and Correction Library for HPC
SESSION: Research Poster Reception
EVENT TYPE: ACM Student Research Competition Poster, Poster, Electronic Poster
TIME: 5:15PM - 7:00PM
SESSION CHAIR: Bernd Mohr
AUTHOR(S):David Fiala, Kurt Ferreira, Frank Mueller, Christian Engelmann
ROOM:WSCC North Galleria 2nd/3rd Floors
ABSTRACT:
Proposed exascale systems will present a number of considerable resiliency challenges. In particular, DRAM soft-errors, or bit-flips, are expected to greatly increase due to the increased memory density of these systems. Current hardware-based fault-tolerance methods will be unsuitable for addressing the expected soft error frequency rate. As a result, additional software will be needed to address this challenge. In this paper we introduce LIBSDC, a tunable, transparent silent data corruption detection and correction library for HPC applications. LIBSDC provides comprehensive SDC protection for program memory by implementing on-demand page integrity verification by utilizing the MMU. Experimental benchmarks with Mantevo HPCCG show that once tuned, LIBSDC is able to achieve SDC protection with less than 100% overhead of resources.
Chair/Author Details:
- Bernd Mohr (Chair) - Juelich Supercomputing Centre
- David Fiala - North Carolina State University
- Kurt Ferreira - Sandia National Laboratories
- Frank Mueller - North Carolina State University
- Christian Engelmann - Oak Ridge National Laboratory
A scalable two-phase parallel I/O library with application to a large scale subsurface simulator
SESSION: Research Poster Reception
EVENT TYPE: ACM Student Research Competition Poster, Poster, Electronic Poster
TIME: 5:15PM - 7:00PM
SESSION CHAIR: Bernd Mohr
AUTHOR(S):Sarat Sreepathi, Vamsi Sripathi, Glenn Hammond, Richard Mills, Kumar Mahinthakumar
ROOM:WSCC North Galleria 2nd/3rd Floors
ABSTRACT:
This poster describes the development of a highly scalable application layer parallel I/O library (ASCEM-IO) for scientific applications. This library was envisioned to leverage our earlier I/O optimization experience to build a scalable general purpose parallel I/O capability for any application. The parallel I/O library provides a higher level API (Application Programming Interface) to read and write large scientific datasets in parallel at very large processor counts. Specifically, the goal is to take advantage of existing parallel I/O libraries, such as HDF5 which are being widely used by scientific applications and modify these algorithms to better scale on larger number of processors. This is accomplished by dividing the traditional I/O operations (read/write) into two phases, a communication phase and an I/O phase. Results with a real application on the Cray XT/5 indicates significant performance improvement on large processor cores when compared to default HDF collective I/O operations.
Chair/Author Details:
- Bernd Mohr (Chair) - Juelich Supercomputing Centre
- Sarat Sreepathi - North Carolina State University
- Vamsi Sripathi - North Carolina State University
- Glenn Hammond - Pacific Northwest National Laboratory
- Richard Mills - Oak Ridge National Laboratory
- Kumar Mahinthakumar - North Carolina State University
Fast One-Sided Communication on Supercomputers and Application to Three Scientific Codes
SESSION: Research Poster Reception
EVENT TYPE: ACM Student Research Competition Poster, Poster, Electronic Poster
TIME: 5:15PM - 7:00PM
SESSION CHAIR: Bernd Mohr
AUTHOR(S):Jeff R. Hammond, Sreeram Potluri, Zheng (Cynthia) Gu, Alex Dickson, James Dinan, Ivo Kabadshow, Pavan Balaji, Vinod Tipparaju
ROOM:WSCC North Galleria 2nd/3rd Floors
ABSTRACT:
We have developed a new library for one-sided communication and applied it to three different scientific codes: the NWChem computational chemistry application, the ScaFaCoS (Scalable Fast Coulomb Solvers) library, and the NEUS (non-equilibrium umbrella sampling) application. All three codes rely upon asynchronous communication, such as one-sided put, get, accumulate and remote atomics (e.g. fetch-and-add). Our library was designed to meet the requirements of current and next-generation interconnects found in Blue Gene/P and Q, PERCS and Cray XE architectures, such as dynamic routing, DMA engines and hardware remote atomics. The synchronization semantics and implementation have been designed for scalability to more than one million ranks. We demonstrate the scaling of all three applications to thousands of cores. As a scaling exemplar, the fast multiple method (FMM) in ScaFaCoS is demonstrated to scale to 300K ranks of Jugene, which is impossible with MPI and ARMCI for both semantic and implementation reasons.
Chair/Author Details:
- Bernd Mohr (Chair) - Juelich Supercomputing Centre
- Jeff R. Hammond - Argonne National Laboratory
- Sreeram Potluri - Ohio State University
- Zheng (Cynthia) Gu - Florida State University
- Alex Dickson - University of Chicago
- James Dinan - Argonne National Laboratory
- Ivo Kabadshow - Juelich Supercomputing Centre
- Pavan Balaji - Argonne National Laboratory
- Vinod Tipparaju - Oak Ridge National Laboratory
ORNL/UT participation in SC2011 BOFs
Tuesday
Open source file systems – transitioning from Petascale to Exascale
SESSION: Open source file systems – transitioning from Petascale to Exascale
EVENT TYPE: Birds of a Feather
TIME: 12:15PM - 1:15PM
SESSION LEADER(S):Galen Shipman, Stephane Thiell
ROOM:TCC 301/302
ABSTRACT:
As the high-performance computing community makes the transition from Petascale to Exascale systems parallel file systems will be required to provide unprecedented levels of performance and capacity. The community developed Exascale roadmap points towards an I/O system capable of supporting 60 Terabytes/sec of aggregate bandwidth and up to an Exabyte of capacity. Dramatic increases in total concurrency, up to one billion in a future Exascale system, will require revolutionary advances in parallel file system software to effectively exploit the underlying I/O system capabilities while supporting this level of concurrency. In this birds of a feather session Open Scalable File Systems and the European Open File Systems Cooperative will organize a panel of experts to discuss the challenges and potential solutions to Exascale I/O. Our panel includes: Eric Barton – Whamcloud (Lustre), Paul Nowoczynski – PSC (ZEST, SLASH2), Rob Ross – ANL (PVFS-2), and Sage Weil - DreamHost (Ceph).
Session Leader Details:
- Galen Shipman (Primary Session Leader) - OpenSFS and ORNL
- Stephane Thiell (Secondary Session Leader) - EOFS and CEA
Scalable Heterogeneous Computing Forum
SESSION: Scalable Heterogeneous Computing Forum
EVENT TYPE: Birds of a Feather
TIME: 12:15PM - 1:15PM
SESSION LEADER(S):Jeffrey Vetter, Sadaf Alam
ROOM:TCC 304
ABSTRACT:
As scalable heterogeneous computing (SHC) architectures are emerging as solutions for HPC, data-intensive and engineering applications, there is an imminent need for a forum that focuses on operational, procurement, and deployment requirements as well as R&D priorities for the SHC ecosystem. This BOF launches an international, open, customer-driven community effort for developing SHC ecosystems solutions and provides attendees an opportunity to join the forum. Speakers from academic, government, and vendor institutions will introduce preliminary working groups, and the organizing committee will present both a draft version of a ‘Best Practices’ guide on SHC clusters and preliminary survey results identifying the communities’ priorities. This forum will collaborate and coordinate with ongoing operational, R&D, and standardization efforts across government, industry, and academia. These efforts include programming environments, system administration, and runtime systems. Furthermore, the forum will identify technical gaps across these communities, and encourage investments in these gaps from member portfolios.
Session Leader Details:
- Jeffrey Vetter (Primary Session Leader) - Oak Ridge National Laboratory and Georgia Tech
- Sadaf Alam (Secondary Session Leader) - Swiss National Supercomputing Centre
The 2011 HPC Challenge Awards
SESSION: The 2011 HPC Challenge Awards
EVENT TYPE: Birds of a Feather
TIME: 12:15PM - 1:15PM
SESSION LEADER(S):Piotr Luszczek, Jeremy Kepner
ROOM:TCC LL2
ABSTRACT:
The 2011 HPC Challenge Awards is a competition with awards in two classes: Class 1: Best Performance awards the best run submitted to the HPC Challenge website. The winners in four categories are announced: HPL, Global-RandomAccess, EP-STREAM-Triad per system, and Global-FFT. Class 2: Most Elegant awards an implementation of three or more of the HPC Challenge benchmarks with special emphasis put on: HPL, Global-RandomAccess, STREAM-Triad and Global-FFT. This award would be weighted 50% on performance and 50% on code elegance/clarity/size. In addition, we allow presentation of additional codes that are critical to scientific applications. Competition in Class 1 offers a rich view of contemporary supercomputers as they compete for supremacy not just in one category but in four. Class 2, on the other hand, offers a glimpse into high end programming technologies and the effectiveness of their implementation. The audience will have a chance to look at new programming environments.
Session Leader Details:
- Piotr Luszczek (Primary Session Leader) - University of Tennessee, Knoxville
- Jeremy Kepner (Secondary Session Leader) - MIT Lincoln Laboratory
Computer Architecture Simulation Framework
SESSION: Computer Architecture Simulation Framework
EVENT TYPE: Birds of a Feather
TIME: 5:30PM - 7:00PM
SESSION LEADER(S):Almadena Y. Chtchelkanova, Stephen W. Poole
ROOM:TCC 205
ABSTRACT:
The goal of this BOF is to engage the academic community, government and industry to make sure that there is a community framework allowing interoperability of simulation tools for Computer Architecture. The framework of interoperable simulators for computer architecture is long overdue. It is important to bring together and build a community of researchers and implementers of architecture simulation, emulation, and modeling tools for memory, CPU/cache, storage and the network on a chip.
Session Leader Details:
- Almadena Y. Chtchelkanova (Primary Session Leader) - National Science Foundation
- Stephen W. Poole (Secondary Session Leader) - ORNL/DoD
Next Generation I/O and visualization requirements for in situ processing
SESSION: Next Generation I/O and visualization requirements for in situ processing
EVENT TYPE: Birds of a Feather
TIME: 5:30PM - 7:00PM
SESSION LEADER(S):Scott Klasky, Hasan Abbasi, Manish Parashar
ROOM: TCC LL1
ABSTRACT:
Next generation exascale systems will require HPC to revolutionize the concept of I/O, data analysis, and visualization. HPC is ready for this revolution, where we can build tools, which are easy-to-use, sustainable, scalable, and portable. This BOF will focus on generating new ideas to design new algorithms and middleware to aid scientist in high performance, data intensive computing. Leadership class computing has allowed application scientists to solve problems that would have been impossible 5 years ago. Much of the focus of the community has gone into creating scalable science to work efficiently on these petascale platforms. One consequence has been that the data generated from these simulations has ballooned to unprecedented volumes, opening up new challenges and opportunities. Simulations of this scale need to incorporate new techniques to meet these challenges which requires tools that can help experts to make real-time decisions of what to do with the data.
Session Leader Details:
- Scott Klasky (Primary Session Leader) - Oak Ridge National Laboratory
- Hasan Abbasi (Secondary Session Leader) - Oak Ridge National Laboratory
- Manish Parashar (Secondary Session Leader) - Rutgers University
Wednesday
Open MPI State of the Union
SESSION: Open MPI State of the Union
EVENT TYPE: Birds of a Feather
TIME: 12:15PM - 1:15PM
SESSION LEADER(S): Jeff Squyres, George Bosilca
ROOM: TCC 303
ABSTRACT:
MPI continues to evolve, both as a standard and as software implementations. The members of the Open MPI project are both driving the evolution of MPI as well as pushing the boundaries of what is possible in extreme computing environments. The K supercomputer validates this approach: a Fujitsu team made local modifications to Open MPI and achieved over 8 petaflops. But at the same time, much effort is put enabling newcomers to HPC to have an MPI that "just works" and is featurized to support both a wide variety of environments. Open MPI community's unique blend of academics, researchers, system administrators, and vendors provide many different viewpoints for what make an MPI implementation successful. Work is ongoing in many areas that are directly applicable to real-world HPC applications and users. Join us at the BOF to hear a state of the Union for Open MPI. New contributors are welcome!
Session Leader Details:
- Jeff Squyres (Primary Session Leader) - Cisco
- George Bosilca (Secondary Session Leader) - University of Tennessee, Knoxville
Scientific Data on the Path to Exascale: Lessons, Insights and Predictions from 10+ years on the front lines
SESSION: OpenSHMEM
EVENT TYPE: Birds of a Feather
TIME: 5:30PM - 7:00PM
SESSION LEADER(S):Tony Curtis, Steve Poole
ROOM:TCC 303
ABSTRACT:
OpenSHMEM: From Message Passing to PGAS the Easy Way OpenSHMEM is the new open community standard for the one-sided PGAS/RMA-based communication paradigm that dominates the solution space for several problem domains that exist on massively parallel systems. We invite users, application developers and researchers from industry, academic and government communities to participate in developing an open organization and specification for OpenSHMEM. The initial specification is based on the original SGI API (SHMEM is a trademark of SGI), and now is the ideal time to join us and engage in the ongoing active extension of the OpenSHMEM specification for version 2.0! Background material, documentation, benchmarks, a V&V test suite, and the reference implementation will be available on http://www.openshmem.org/ prior to SC11.
Session Leader Details:
- Tony Curtis (Primary Session Leader) - University of Houston
- Steve Poole (Secondary Session Leader) – ORNL
Petascale Systems Management BOF
SESSION: Petascale Systems Management BOF
EVENT TYPE: Birds of a Feather
TIME: 5:30PM - 7:00PM
SESSION LEADER(S):William R. Scullin, Narayan Desai, Matt Ezell
ROOM:WSCC 204
ABSTRACT:
Bringing together a wide range of systems administrators, the Petascale Systems management BOF aims to: identify the issues that commonly arise with the management of petascale systems, such as Blue Gene /P, Cray XT5 systems, and extremely large clusters; share solutions and approaches to issues; introduce members of the systems administration community to one another.
Session Leader Details:
- William R. Scullin (Primary Session Leader) - Argonne National Laboratory
- Narayan Desai (Secondary Session Leader) - Argonne National Laboratory
- Matt Ezell (Secondary Session Leader) - University of Tennessee, Knoxville
The IESP and EESI Efforts
SESSION: The IESP and EESI Efforts
EVENT TYPE: Birds of a Feather
TIME: 5:30PM - 7:00PM
SESSION LEADER(S):Jack Dongarra, Pete Beckman, Jean-Yves Berthou
ROOM:TCC 304
ABSTRACT:
Over the last twenty years, the international open source community has increasingly provided the software at the heart of the world’s high-performance computing systems. The community provides everything from operating system components to compilers and advanced math libraries. As an international community we have only loosely coordinated activities and plans for development. The new rapidly changing technologies in multicore, power consumption, GPGPUs, and memory architectures creates an opportunity for the community to work together and build an international program to design, build, and deliver the software so critical the science goals of our institutions. To help plan how the international community could build a partnership to provide the next generation of HPC software to support scientific discovery, we have had a series of workshops. We plan to report on the effort and solicit feedback from the community. This BOF will also cover the European Exascale Software Initiative (EESI).
Session Leader Details:
- Jack Dongarra (Primary Session Leader) - University of Tennessee, Knoxville
- Pete Beckman (Secondary Session Leader) - Argonne National Laboratory
- Jean-Yves Berthou (Secondary Session Leader) – EDF
The MPI 3.0 Standard - The Final Stages
SESSION: The MPI 3.0 Standard - The Final Stages
EVENT TYPE: Birds of a Feather
TIME: 5:30PM - 7:00PM
SESSION LEADER(S):Richard Graham
ROOM:TCC 301/302
ABSTRACT:
The Message Passing Interface (MPI) is the ubiquitous standard used by parallel scientific applications to satisfy their communication and process control needs(http://www.mpi-forum.org ). The MPI Forum, which is the MPI standards body, is in the last stages of finalizing version 3.0 of the standard, with several functional enhancements on their way to final standardization. Support for several things has already been added to the MPI 3.0 draft standard, including nonblocking collective, with features, such as support for MPI fault-tolerance, improved support for remote memory operations, and standardized hooks for tool access to information internal to MPI implementations, are in the in final stages of standardization. We will discuss the status of the MPI-3 standardization effort and solicit input from attendees aspects of the standard that should be improved past the MPI 3.0 standard.
Session Leader Details:
- Richard Graham (Primary Session Leader) - Oak Ridge National Laboratory
Thursday
Developing, Recruiting, and Retaining a Diverse Workforce in HPC
SESSION: Developing, Recruiting, and Retaining a Diverse Workforce in HPC
EVENT TYPE: Birds of a Feather
TIME: 12:15PM - 1:15PM
SESSION LEADER(S):Rebecca Hartman-Baker, Samantha Foley, Judith Hill
ROOM:TCC LL2
ABSTRACT:
While the majority of people are in principle supportive of efforts to increase the population of underrepresented groups in high-performance computing, few know how to do so in practice. In this BOF we will present concrete advice from people who have been successful at recruiting and retaining students, postdocs, and employees from diverse backgrounds. We will begin with a discussion of the challenges that underrepresented minorities face in the HPC workplace, including challenges unique to the field, and examine practices that can be implemented to alleviate these challenges. The panelists, representing academic, industrial, and laboratory HPC workplaces, will present ideas and strategies that have met with success. The session will continue with a Q&A session and an open discussion. In this BOF, we hope to empower employers and supervisors in the HPC field with the tools to create a work or educational environment where people of all types can flourish.
Session Leader Details:
- Rebecca Hartman-Baker (Primary Session Leader) - Oak Ridge National Laboratory
- Samantha Foley (Secondary Session Leader) - Oak Ridge National Laboratory
- Judith Hill (Secondary Session Leader) - Oak Ridge National Laboratory
Heterogeneous supercomputing on ORNL's Titan
SESSION: Heterogeneous supercomputing on ORNL's Titan
EVENT TYPE: Birds of a Feather
TIME: 12:15PM - 1:15PM
SESSION LEADER(S):Adam B. Simpson, Anthony DiGirolamo
ROOM:WSCC 3B
ABSTRACT:
The Oak Ridge Leadership Computing Facility's (OLCF) next leadership-class system, Titan, will thrust heterogeneous computing to the forefront of supercomputing and bring along with it a new set of programming challenges. This BoF will provide an introduction to the specific tools and techniques that the OLCF plans to offer in order to address these challenges and then transition into an open discussion. During these discussions, participants are encouraged to share their difficulties or expected difficulties with porting their scientific code to a massively hybrid system like Titan. The intent of this BoF is to provide an avenue to discuss specifically what the OLCF can provide to ease the transition into heterogeneous computing on Titan.
Session Leader Details:
- Adam B. Simpson (Primary Session Leader) - ORNL
- Anthony DiGirolamo (Secondary Session Leader) – ORNL
HPC Centers
SESSION: HPC Centers
EVENT TYPE: Birds of a Feather
TIME: 5:30PM - 7:00PM
SESSION LEADER(S): Robert Whitten Jr
ROOM:TCC LL2
ABSTRACT:
This BOF provides an open forum for user assistance personnel to discuss topics of interest. Topics include but are not limited to ticket procedures, queue policies, organization and structure of support options, and any pressing topic. The goal of this BoF is to provide an opportunity for user support staff from HPC centers (NSF, DOE, DOD, NASA, university, and others from USA and around the world) to share ideas and discuss common problems and concerns. Members should expect to share, discuss and compare Best Practices in order to further the goals of HPC centers. This BOF is an ongoing meeting of the HPC Centers working group and all interested parties are encouraged to attend.
Session Leader Details:
- Robert Whitten Jr (Primary Session Leader) - Oak Ridge National Laboratory
Women in HPC: Community Building
SESSION: Women in HPC: Community Building
EVENT TYPE: Birds of a Feather
TIME: 5:30PM - 7:00PM
SESSION LEADER(S):Samantha S. Foley, Rebecca J. Hartman-Baker, Judith C. Hill
ROOM:WSCC 613/614
ABSTRACT:
Despite the fact that women comprise more than half the world's population, they are underrepresented in high-performance computing. One way to improve gender diversity in HPC is to build communities for women to interact, advise and share success stories to inspire and encourage women at all career levels. We will begin with a presentation on the goals of the session, and information on creating and maintaining a strong community. Participants will then form groups to brainstorm ideas for building communities for women in HPC in local and online settings, as well as at SC. The remainder of the time will be spent discussing the ideas generated in the brainstorming activity, planning ways to implement them, and general discussion related to supporting women in HPC. We hope that attendees will gain new friends and colleagues, and generate more opportunities to interact throughout the year.
Session Leader Details:
- Samantha S. Foley (Primary Session Leader) - Oak Ridge National Laboratory
- Rebecca J. Hartman-Baker (Secondary Session Leader) - Oak Ridge National Laboratory
- Judith C. Hill (Secondary Session Leader) - Oak Ridge National Laboratory
ORNL/UT participation in SC2011 Workshops
Monday
2nd SC Workshop on Petascale Data Analytics: Challenges and Opportunities
SESSION: 2nd SC Workshop on Petascale Data Analytics: Challenges and Opportunities
EVENT TYPE: Workshop
TIME: 9:00AM - 5:30PM
Organizer(s):Ranga Raju Vatsavai, Scott A. Klasky, Vipin Kumar
ROOM:Hyatt Leonesa II
ABSTRACT:
Recent decade has witnessed data explosion, and petabyte sized data archives are not uncommon any more. Many traditional application domains are now becoming data intensive. It is estimated that organizations with high end computing infrastructures and data centers are doubling the amount of data that they are archiving every year. Recent trends show that the cloud computing is becoming a more practical and economical solution for many organizations. Cloud computing has attracted significant attention of both industry and academia in recent years. The first edition of this workshop organized with SC-2010 has attracted more than 300 registrations, which demonstrates the popularity of this topic among SC attendees. In addition to the cloud focus, this years workshop will also look into the emerging middleware and software services that are specifically aimed at facilitating extreme scale data analytics.
Chair/Organizer Details:
- Ranga Raju Vatsavai - Oak Ridge National Laboratory
- Scott A. Klasky - Oak Ridge National Laboratory
- Vipin Kumar - University of Minnesota
Computational Methods for Biosurveillance
SESSION: Computational Methods for Biosurveillance
EVENT TYPE: Workshop
TIME: 9:00AM - 5:30PM
Organizer(s):Robert Cottingham, Richard Stouder
ROOM:Hyatt Leonesa III
ABSTRACT:
Biosurveillance has been primarily based on PCR detection, and intelligence gathering and analysis. With the increasing potential for terrorist engineered threats, and the computational means to detect both natural and synthetic threats, this workshop brings together experts in related capabilities, both biological and computational, to consider existing and proposed methods for computational biosurveillance. The workshop aims to provide an integrated view of the state of the art, identify technical challenges, and develop a vision for the future of computational biosurveillance. Relevant topics include data intensive computing for analysis of multiple real-time data streams, genomic and transcriptomic analysis of environmental samples, social network analysis, sensor network integration, spatial analysis and visualization, agent-based simulation.
Chair/Organizer Details:
- Robert Cottingham - Oak Ridge National Laboratory
- Richard Stouder - Oak Ridge National Laboratory
Scalable Algorithms for Large-Scale Systems
SESSION: Scalable Algorithms for Large-Scale Systems
EVENT TYPE: Workshop
TIME: 9:00AM - 5:30PM
Organizer(s):Vasil Alexandrov, Al Geist, Jack Dongarra, Christian Engelmann
ROOM:Hyatt Princessa II
ABSTRACT:
Novel scalable scientific algorithms are needed in order to enable key science applications to exploit the computational power of large-scale systems. This is especially true for the current tier of leading petascale machines and the road to exascale computing as HPC systems continue to scale up compute nodes and processor core count. These extreme-scale systems require novel scientific algorithms to hide network and memory latency, have very high computation/communication overlap, have minimal communication, and have no synchronization points. Scientific algorithms for multi-petaflop and exa-flop systems also need to be fault tolerant and fault resilient, since the probability of faults increases with scale. Resilience at the system software and at the algorithmic level is needed as a crosscutting effort. Key science applications require novel mathematical models and system software that address the scalability and resilience challenges of current- and future-generation extreme-scale HPC systems.
Chair/Organizer Details:
- Vasil Alexandrov - Barcelona Supercomputing Center
- Al Geist - Oak Ridge National Laboratory
- Jack Dongarra - University of Tennessee, Knoxville
- Christian Engelmann - Oak Ridge National Laboratory
ORNL/UT participation in SC2011 Related Activities
Jaguar & Spider storage system, and the latest initiative: Titan.WHEN: Monday, November 14th, 11:30am - 2pm
WHERE: Sheraton Hotel Diamond Room (Note: Sheraton Hotel is adjacent to convention center)
Lunch will start at 11:30, with presentations starting at noon
Keynote Speakers:
- Introductions from John Josephakis (SVP of HPC/LS Worldwide Sales)
- Mark Seager, Chief Technology Officer for the HPC Ecosystem, Intel: Intel's vision for HPC in the Exascale era
- Jeff Denworth, VP of Marketing for DDN: Latest product announcements, key issues for I/O on the path to Exascale.
- 1:00 - 2:00 Galen Shipman, Technology Integration Leader for Oak Ridge National Laboratories: Jaguar & Spider storage system, and the latest initiative: Titan.
| Speaker | Presentation Title | Job Title | Affiliation | |
MONDAY |
||||
| Jack Dongarra | The November 2011 TOP500 List | Distinguished Professor of Computer Science | University of Tennessee and Oak Ridge National Laboratory | |
TUESDAY |
||||
| 10:30am | Jim Rogers | OLCF Director of Operations | Oak Ridge National Laboratory | |
| 11:30am | Doug Kothe | CASL - The Consortium for Advanced Simulation of Light Water Reactors: A DOE Energy Innovation Hub for Modeling and Simulation of Nuclear Reactors | Director of CASL | Oak Ridge National Laboratory |
| 1:30pm | Bland or Wells | OLCF Project Director/OLCF Director of Science | Oak Ridge National Laboratory | |
| 2:30pm | Bronson Messer | The Center for Accelerated Application Readiness: Uncovering Hierachical Parrellism | Senior R&D Staff | Oak Ridge National Laboratory |
| 3:30pm | Galen Shipman | Group Leader, Tech Integration | Oak Ridge National Laboratory | |
WEDNESDAY |
||||
| 10am | Joseph Oefelein | Large Eddy Simulation of Combustion for Propulsion and Power: Experiences on the OLCF Cray-XT Jaguar | Distinguished Member of Technical Staff | Sandia National Laboratory |
| 11am | Sean Ahern | XSEDE: The eXtreme Science and Engineering Discovery Environment | NICS Executive Director | Oak Ridge National Laboratory |
| 1pm | Rich Graham | Titan's programming environment (PE) : The challenge of heterogeneity | Group Leader, Application Performance Tools | Oak Ridge National Laboratory |
| 2pm | Marios Soteriou | HPC as a Key Enabler for High Fidelity Simulation of Sprays in Aerospace Applications | Fellow of Research, Thermo-fluid Sciences | United Technologies Research Center |
| 3pm | Daniel Quest | Computational Methods for Biosurveillance | Group Leader, Computational Biology & Bioinformatics | Oak Ridge National Laboratory |
THURSDAY |
||||
| 10am | Les Button | HPC and Innovation: A View from "The Missing Middle" | Research Fellow, Science & Technoloogy | Corning, Inc. |
| 11:30am | Cecilia Bitz (NICS User) | Antarctic climate change caused by ozone depletion in a fine-resolution ocean climate model | Associate Professor | University of Washinton |
| Author(s) | Title | File |
| Sean Ahern, Jian Huang, Wes Bethel, Scott Klasky, Dave Semeraro, George Ostrouchov, Miron Livny | RDAV Center for Remote Data Analysis and Visualization | |
| Rommie Amaro, Adrian Roitberg, and Ross C. Walker | Simulations of Influenza Neuraminidase In Search of the Next Generation of Flu Drugs | |
| Rick Archibald, Ralf Deiterding, Cory Hauck, Dongbin Xiu | Advanced Dynamically Adaptive Algorithms for Stochastic Simulations on Extreme Scales | |
| Barbara Beckerman | Biomedical Science and Engineering at Oak Ridge National Laboratory | |
| Budhendra Bhaduri, Alexandre Sorokine, Cheng Liu | High-Performance Visualization of Geographic Data | |
| Budhendra Bhaduri | LandScan Population Research Program | |
| Cecilia Bitz | How does Earth's Climate Respond to Ozone Depletion | |
| Buddy Bland | Jaguar: Petascale Science Delivered | |
| George Bosilca | Dealing with the Scale Challenge | |
| Glenn Brook | Application Acceleration Center of Excellence | |
| Adam Burrows and Jason Nordhaus | 3D Core-collapse Supernova Simulations | |
| Kate Carter | High Performance Recruiting Changing the World One Career at a Time | |
| Barbara Chapman, Tony Curtis, Charles Koelbel, Jeffery Kuehn, Stephen Poole, and Lauren Smith | Introducing OpenSHMEM | |
| Jacqueline H. Chen (PI), Chun Sang Yoo, Gaurav Bansal, Hongfeng Yu, Ajith Mascarenhas, Ray Grout, Tianfeng Lu, and Ramanan Sankaran | High Fidelity Direct Numerical Simulations of Turbulent Combustion | |
| Charlotte Christensen, Tom Quinn, Fabio Governato, and the N-Body Shop | Simulating the Cold Molecular Hydrogen Gas in Dwarf Galaxies | |
| John Cobb | DataONE: Enabling Data-Intensive Biological and Environmental Research through Cyberinfrastructure | |
| John Cobb | FACETS: Framework Application for Core-Edge Transport Simulations | |
| David C. Collins, Paolo Padoan, Michael L. Norman, and Hao Xu | Star Formation with Turbulence and Magnetic Fields | |
| Lonnie D. Crosby, Justin L. Whitt, Patricia A. Kovatch, George Herring, and Chris Buckley | A Regression Testing Framework for Non-catastrophic Failure Detection | |
| Ed D’Azevedo | Computational Mathematics at the Oak Ridge National Laboratory | |
| Eva Darian | Structural Mechanism Associated with Domain Opening in GOF Mutations in SHP2 Phosphatase | |
| Frank DeNap | SensorNet: The New Science of Public Protection and Awareness | |
| Tiziana Di Matteo | Simulating the Universe on a Supercomputer | |
| Jack Dongarra, Shirley Moore, and Dan Terpstra | High-Productivity Software Development Tools | |
| Jack Dongarra, Heike Jagode, Shirley Moore, and Dan Terpstra | Component PAPI: Performance Measurement Beyond the CPU | |
| Sudip S. Dosanjh and Jeff Nichols | Institute for Advanced Architectures and Algorithms: Overview and Projects | |
| Innovative Computing Laboratory - University of Tennessee, University of California–Berkeley, and University of Colorado–Denver |
PLASMA Parallel Linear Algebra Software for Multicore Architectures | |
| J. Engel, J. Terasaki, and M. Mustonen | Nuclear Density Functional Theory for Excitations | |
| Matthew Ezell, Ryan Braby, and Patricia Kovatch | The Malthusian Catastrophe is Upon Us! Are the Largest HPC Machines Ever Up? | |
| Mark Fahey | Petascale Hardware at NICS | |
| Mark Fahey | Kraken - NSF's Petascale Resources at NICS | |
| James W. Ferguson | Education Outreach and Training | |
| Miguel Fuentes-Cabrera | Theoretical Studies of the De-wetting Mechanism of Metallic Thin-Films | |
| Auroop R. Ganguly | Climate Extremes: The Science, Impacts, and Policy Relevance | |
| Huajian Gao | Probing Mechanical Properties of Nanostructured Materials via Large Scale Molecular Dynamics Simulations | |
| Al Geist | Fault-Tolerance Challenges and Solutions | |
| Richard L. Graham, Pavel Shamis, Joshua S. Ladd, and Manjunath Gorentla Venkata | Cheetah: A Framework for Scalable Collective Operations | |
| Richard L. Graham, Pavel Shamis, Joshua S. Ladd, Manjunath Gorentla Venkata, Noam Bloch, and Gilad Shainer | Network Offloaded Hierarchical Collectives Using ConnectX-2’s CORE-Direct | |
| L. Grinberg, G. E. Karniadakis, D. Fedosov, FZ Juelich J. A. Insley, V. Morozov, and M. E. Papka | Human Arterial-Tree Multi-scale Modeling, Simulations and Visualization | |
| Bilel Hadri and Mark Fahey | Automatic Library Tracking Database for High Performance Computing Systems | |
| Bilel Hadri and Haihang You | Performance Comparison Framework for Numerical Libraries on Cray XT5 Systems | |
| G. Hagen, H. Nam, W. Nazarewicz, and T. Papenbrock | Building the Atomic Nucleus from the Ground Up | |
| Forrest M. Hoffman, James T. Randerson, Peter E. Thornton, Natalie M. Mahowald, Keith Lindsay, Yen-Huei Lee, Cynthia D. Nevison, Scott C. Doney, Gordon B. Bonan, Reto Stöckli, Curtis C. Covey, Steven W. Running, and Inez Y. Fung | Systematic Assessment of Terrestrial Biogeochemistry in Coupled Climate–Carbon Models | |
| Terry Jones (ORNL), Laxmikant Kalé (UIUC) Celso Mendes (UIUC), Esteban Meneses (UIUC) Yanhua Sun (UIUC), José Moreira (IBM) Eliezer Dekel (IBM), Gennady Laventman (IBM) Yoav Tock (IBM), Benjamin Mandler (IBM) | Adaptable System Software for A Brave New World of Revolutionary Computer Architectures | |
| Homa Karimabadi | Breakthrough Kinetic Simulations of the Earth's Magnetosphere | |
| Kiran Kumar Kasichayanula, Haihang You, Shirley Moore, Stanimire Tomov, Heike Jagode, and Matt Johnson | Power-aware Computing on GPGPUs | |
| Ricky A. Kendall | Scientific Computing Group | |
| Scott Klasky | End-to-End Computing at ORNL: Preparing for the Exascale | |
| Scott Klasky | Extreme Scale Computing with the Fusion Particle Code XGC1 | |
| Fanyou Kong, Ming Xue, and Kevin Thomas | CAPS 4 km Ensemble and 1 km Deterministic Forecast for NOAA HWT | |
| Douglas B. Kothe | CASL: The Consortium for Advanced Simulation of Light Water Reactors | |
| Alexei G. Kritsuk | From Molecular Clouds to Protostars | |
| Chee Wai Lee | TAU Demonstration | |
| Seyong Lee, Jeffrey Vetter, and Rudolf Eigenmann | OpenMPC: OpenMP Extended for CUDA | |
| Kalev Leetaru | The GlobalNet Project: Forecasting Human Society Through Petascale News Analysis | |
| Anyi Li | Understanding the Beginning of the Universe | |
| Anthony Danalis and Jack Dongarra | DARPA AACE Blackjack | |
| Piotr Luszczek | The HPC Challenge (HPCC) Benchmark Suite | |
| Barney Maccabe | The Hybrid Multicore Consortium (HMC) | |
| Thomas Maier | Pairing in multi-Fermi-surface systems: New insight into unconventional superconductivity with advanced computing | |
| Debbie McCoy | Research Alliance in Math and Science | |
| Collin McCurdy and Jeffrey S. Vetter | Memphis: Finding and Fixing NUMA-Related Performance Problems on Multi-core Platforms | |
| Amy McGovern | Understanding the Formation of Tornadoes Through Spatiotemporal Relational Data Mining | |
| Jeremy Meredith and Kyle Spafford | Accelerating Scientific Applications with GPUs | |
| Bronson Messer | OLCF Science Overview | |
| Anthony Mezzacappa | Toward the Explosion Mechanism for Core-Collapse Supernovae: An Emerging Picture | |
| Richard Tran Mills | SciDAC-2 Project: Modeling Multiscale-Multiphase-Multicomponent Subsurface Reactive Flows Using Advanced Computing INCITE Project: Modeling Reactive Flows in Porous Media | |
| Shirley Moore and Dan Terpstra | Multiple Metrics Modeling Infrastructure (MuMMI) | |
| Jeff Nichols | ORNL Collaborators in HPC | |
| Phani Nukala | Statistical Physics of Fracture: Recent Advances through High-Performance Computing | |
| Joseph C. Oefelein, (PI) Guilhem Lacaze, Rainer Dahms, and Ramanan Sankaran | High-Fidelity Simulations for Advanced Engine Combustion Research | |
| Meers Oppenheim, Yakov Dimant, and Yann Tambouret | Plasma Instabilities Generated in the Ionosphere | |
| George Ostrouchov | Statistics at Scale | |
| Sreekanth Pannala | Micro-mesoscopic Modeling of Heterogeneous Chemically Reacting Flows (MMMHCRF) over Catalytic/Solid Surfaces | |
| Daniel Pelfrey | NCCS Network Roadmap | |
| Kalyan S. Perumalla | Parallel Discrete Event Simulation (PDES) at ORNL | |
| Thomas E. Potok | Novel, Large Scale, High Speed Text and Data Analysis Research | |
| Dave Pugmire | Visualization at the Oak Ridge Leadership Computing Facility | |
| Nagi Rao | Information Fusion: Science and Engineering of Combining information from Multiple Sources | |
| Nagi Rao | DOE UltraScience Net: High-Performance Experimental Network Research Testbed | |
| Bhanu Rekepalli | Systems Biological Workbench Development on NICS Resources | |
| Matthias Rempel and Mark Cheung | Simulations of Solar Active Regions | |
| Philip C. Roth | Performance Engineering Research Institute (PERI) | |
| Nagiza F. Samatova and Arie Shoshani (PI) | Scientific Data Management Center | |
| Christian E. Schafmeister | Applications of Bis-peptide nanostructures | |
| Stephen L. Scott and Christian Engelmann | Reliability, Availability, and Serviceability (RAS) for High-Performance Computing | |
| Stephen L. Scott, Thomas Naughton, and Geoffroy Vallée | System-Level Virtualization and OSCAR-V | |
| The ESG-CET Team | The Earth System Grid (ESG): Turning Climate Datasets into Community Resources | |
| Galen M. Shipman | Petascale System Infrastructure | |
| Jeremy C. Smith | Biosystems | |
| Cristiana Stan and James Kinter | An International, Dedicated High-end Computing Project to Revolutionize Climate Modeling | |
| Richard Stouder | The Computational Methods for Biosurveillance Initiative Shaping the Future Together | |
| Blair D. Sullivan and Chris Groër | Solving Large Graph Problems Using Tree Decompositions: A Computational Study | |
| Suzy Tichenor | Oak Ridge Industrial HPC Partnerships Program | |
| Ranga Raju Vatsavai and Budhendra Bhaduri | Geospatial Data Sciences | |
| Sudharshan Vazhkudai | Robust Storage Management in the Machine Room and Beyond | |
| Jeffrey Vetter, Dong Li, Gabriel Marin, Collin McCurdy, Robert Schreiber, Trevor Mudge, and Yuan Xie | Blackcomb Project: Year One | |
| Jeffrey S. Vetter, Weikuan Yu, Yong Chen, and Philip C. Roth | Understanding and Optimizing Data Input/Output of Large-Scale Scientific Applications | |
| Jeffrey Vetter, Jack Dongarra, Richard Fujimoto, Thomas Schulthess, Karsten Schwan, Sudha Yalamanchili, Richard Glassbrook, Jim Ferguson, Patricia Kovatch, Stephen McNally, Bruce Loftis, Jeremy Meredith, Philip Roth, Kyle Spafford, Jim Rogers, Arlene Washington, and others | Keenland - Enabling Heterogeneous Computing for the Open Science Community | |
| Jeffrey Vetter (PI), Wen-mei Hwu, Allen D. Malony, and Rich Vuduc | Vancouver: A Software Stack for Productive Heterogeneous Exascale Computing | |
| Scott Wells | MAGMA: Matrix Algebra on GPU and Multicore Architectures | |
| Patrick H. Worley | Performance Evaluation and Analysis Consortium (PEAC) End Station | |
| Henian Xia, Kwai Wong, Wenjun Ying, and Xiaopeng Zhao | Numerical Simulation of a Coupled Electro-Mechanical Heart Model |
 Office of Science |
Oak Ridge National Laboratory |
 Department of Energy |