Salishan 2016 Program
Full Program | At-A-Glance Program
Keynote Address
Supercomputers and Superintelligence | Horst Simon, Lawrence Berkeley National Laboratory
Session 1: Hardware Architecture Data Movement Capabilities
Fabric Data Movement: An Integrated Hardware/Software Challenge | Keith Underwood, Intel
Exascale Starts with the Memory System | Bruce Jacob, University of Maryland
Is Memory Scaling Really Over? Myths, Facts, and a Path Forward | Hillery Hunter, IBM
Extreme Scale Computing with Optically Enabled Data Movement | Keren Bergman,, Columbia University
Exascale Starts with the Memory System | Bruce Jacob, University of Maryland
Is Memory Scaling Really Over? Myths, Facts, and a Path Forward | Hillery Hunter, IBM
Extreme Scale Computing with Optically Enabled Data Movement | Keren Bergman,, Columbia University
Session 2: System Software Data Movement Capabilities
Rearchitecting Systems Software for Memory Heterogeneity and Scale | Ada Gavrilovska, Georgia Tech
Data Movement with MPI in a Many-Threaded World| Ryan Grant, Sandia National Laboratories
Active RDMA - New Tricks for an Old Dog | Torsten Hoefler, ETHZ
Got Burst Buffer. Now What? | Adam Moody, Lawrence Livermore National Laboratory
Data Movement with MPI in a Many-Threaded World| Ryan Grant, Sandia National Laboratories
Active RDMA - New Tricks for an Old Dog | Torsten Hoefler, ETHZ
Got Burst Buffer. Now What? | Adam Moody, Lawrence Livermore National Laboratory
Working Dinner
How Facebook Uses Advanced Interconnect Technology to Make the World More Open and Connected | Katharine Schmidtke, Facebook
Session 3: Data Movement from the Applications' Perspective
An Early Start with Multi-Level Memory - Opportunity for Performance or Just the Next Chapter in Programmer Hell? | Si Hammond, Los Sandia National Laboratories
Integrated Modeling for Rapid Assessment and Performance Prediction of HPC Applications | Jason Liu, Florida International University
Data Movement Challenges at NTS: A Paradigm for HPC Applications I/O | Mark Miller, Lawrence Livermore National Laboratory
Data Movement from he Application Perspective | Sanjay Padhi, Amazon
Integrated Modeling for Rapid Assessment and Performance Prediction of HPC Applications | Jason Liu, Florida International University
Data Movement Challenges at NTS: A Paradigm for HPC Applications I/O | Mark Miller, Lawrence Livermore National Laboratory
Data Movement from he Application Perspective | Sanjay Padhi, Amazon
Random Access
Toto, I've a Feeling We're Not on Moore's Law Anymore | Dileep Bhandarkar, Qualcomm
CEA's Data Centric Strategy | Guillaume Colin de Verdiere, CEA
A Standard for Batching BLAS Operations | Jack Dongarra, University of Tennessee
It's About SW Now, Not HW - It's About SW Now, Not Later | Mike Heroux, Sandia National Laboratories
A HYDRA Ensemble Workflow on Trinity | Steven Langer, Lawrence Livermore National Laboratory
Variability: A Performance Nightmare | Allan Porterfield, RENCI
HPX-5 Runtime System Overhead Times | Thomas Sterling, Indiana University
In-Situ Data Analysis of Protein-Folding Trajectories | Michela Taufer, University of Delaware
Programming NVM Systems | Jeffrey Vetter, Oak Ridge National Laboratory
CEA's Data Centric Strategy | Guillaume Colin de Verdiere, CEA
A Standard for Batching BLAS Operations | Jack Dongarra, University of Tennessee
It's About SW Now, Not HW - It's About SW Now, Not Later | Mike Heroux, Sandia National Laboratories
A HYDRA Ensemble Workflow on Trinity | Steven Langer, Lawrence Livermore National Laboratory
Variability: A Performance Nightmare | Allan Porterfield, RENCI
HPX-5 Runtime System Overhead Times | Thomas Sterling, Indiana University
In-Situ Data Analysis of Protein-Folding Trajectories | Michela Taufer, University of Delaware
Programming NVM Systems | Jeffrey Vetter, Oak Ridge National Laboratory
Student Posters
Analyzing Large Data Using In-Situ Visualization and Computer Vision Techniques | Divya Banesh, University of California, Davis
Combining Disparate Data Sources in the HPC Ecosystem | Afredo Gimenez, University of California, Davis
Network-Induced Memory Contention | Taylor Groves, University of New Mexico
Surveying On-Node Task Runtimes Towards a New Tasking Mini-App | Jacob Hemstad, University of Minnesota
Scalable I/O-Aware Job Scheduling for Burst Buffer Enabled HPC Clusters | Stephen Herbein, University of Delaware
Memory & Network Design for Exascale Data Movement | Shang Li, University of Maryland
Task-Based Parallel Computation of the Density Matrix in Quantum-Based Molecular Dynamics Using Graph-Partitioning | Sergio Pino, University of Delaware
Remote Visual Analysis on Large Turbulence Databases at Multiple Scales | Jesus Pulido, University of California, Davis
I/O Aware Power Shifting | Lee Savoie, University of Arizona
Combining Disparate Data Sources in the HPC Ecosystem | Afredo Gimenez, University of California, Davis
Network-Induced Memory Contention | Taylor Groves, University of New Mexico
Surveying On-Node Task Runtimes Towards a New Tasking Mini-App | Jacob Hemstad, University of Minnesota
Scalable I/O-Aware Job Scheduling for Burst Buffer Enabled HPC Clusters | Stephen Herbein, University of Delaware
Memory & Network Design for Exascale Data Movement | Shang Li, University of Maryland
Task-Based Parallel Computation of the Density Matrix in Quantum-Based Molecular Dynamics Using Graph-Partitioning | Sergio Pino, University of Delaware
Remote Visual Analysis on Large Turbulence Databases at Multiple Scales | Jesus Pulido, University of California, Davis
I/O Aware Power Shifting | Lee Savoie, University of Arizona
Session 4: Data Movement from Data Analysis, Workflow, and Visualization Perspectives
Workflow Analysis - An Approach to Characterize Application and System Needs | David Montoya, Los Alamos National Laboratory
Open, Reproducible HPC Data Analysis and Visualization: Minimizing Data Movement | Marcus Hanwell, Kitware
What HPC Can Learn from the Cloud Approach to Big Data | Dale Southard, NVIDIA
Extreme Data Management Analysis and Visualization for Exascale Supercomputers | Valerio Pascucci, University of Utah
Open, Reproducible HPC Data Analysis and Visualization: Minimizing Data Movement | Marcus Hanwell, Kitware
What HPC Can Learn from the Cloud Approach to Big Data | Dale Southard, NVIDIA
Extreme Data Management Analysis and Visualization for Exascale Supercomputers | Valerio Pascucci, University of Utah
Session 5: Input/Output, File Systems, and Data Storage Data Movement Challenges
HPC Storage and I/O Trends and Workflows | Gary Grider, Los Alamos National Laboratory
Accelerating Science with the NERSC Burst Buffer Early User Program | Deborah Bard, NERSC
From File System to Services: Changing the Data Movement Model in HPC| Rob Ross, Argonne National Laboratory
Breaking Free from Globally Serializable OLTP Embedded in Parallel File Systems | Garth Gibson, Carnegie Mellon University
Accelerating Science with the NERSC Burst Buffer Early User Program | Deborah Bard, NERSC
From File System to Services: Changing the Data Movement Model in HPC| Rob Ross, Argonne National Laboratory
Breaking Free from Globally Serializable OLTP Embedded in Parallel File Systems | Garth Gibson, Carnegie Mellon University