Salishan 2019 Program
Full Program
Keynote Address
From Extreme Heterogeneity to Artificial Intelligent Systems | Dimitri Kusnezov, Department of Energy
Session 1: Hardware: Driving and Harnessing Extreme Heterogeneity
3D Integration and Highly Interconnected Systems | Robert Patti, NHanced Semiconductors, Inc.
Nantera NRAM Defines a New Category of "Memory Class Storage" | Bill Gervasi, Nantero
Heterogeneous Accelerators of the Memory, by the Memory and for the Memory | Arun Rodrigues, Sandia National Laboratories
Next Gen Memory | Wendy Elsasser, Arm
Nantera NRAM Defines a New Category of "Memory Class Storage" | Bill Gervasi, Nantero
Heterogeneous Accelerators of the Memory, by the Memory and for the Memory | Arun Rodrigues, Sandia National Laboratories
Next Gen Memory | Wendy Elsasser, Arm
Session 2: Storage Challenges in an Era of Extreme Heterogeneity - As If We Need More I/O Complexity!
Persistent Memory for High-Performance Applications | Tom Coughlin, Coughlin Associates
Accelerating Scientific Discovery with Key-Value Storage | Brad Settlemyer, Los Alamos National Laboratory
Heterogeneity, Schmeterogeneity. Object, Schmobject. Long Live POSIX | John Bent, DDN Storage
Persistence in a Heterogeneous World | Paolo Faraboschi, Hewlett Packard Enterprise
Accelerating Scientific Discovery with Key-Value Storage | Brad Settlemyer, Los Alamos National Laboratory
Heterogeneity, Schmeterogeneity. Object, Schmobject. Long Live POSIX | John Bent, DDN Storage
Persistence in a Heterogeneous World | Paolo Faraboschi, Hewlett Packard Enterprise
Working Dinner
Session 3: Software: Maneuvering Through the Space of Extreme Heterogeneity
Compilers Must Be Fast and So FPGAs Will Never Work! Maybe CGRAs Will Save Us? | Hal Finkel, Argonne National Laboratory
SUSE High Performance Computing - It Just Keeps Getting Better | Jay Kruemcke, SUSE
A Software Architecture for Future Multi-Physics Applications | David Daniel, Los Alamos National Laboratory
HPC Performance Tools: Myths and Facts | Bernd Mohr, Juelich Supercomputing Centre
SUSE High Performance Computing - It Just Keeps Getting Better | Jay Kruemcke, SUSE
A Software Architecture for Future Multi-Physics Applications | David Daniel, Los Alamos National Laboratory
HPC Performance Tools: Myths and Facts | Bernd Mohr, Juelich Supercomputing Centre
Random Access
SW defined HW to improve system efficiency (Virtual Heterogeneity) | Shekhar Borkar, Qualcomm
Experiences putting Tools into Production Codes | Olga Pearce, Lawrence Livermore National Laboratory
Do Androids Dream of Electric Sheep?, and how an Ethical AI Strategy could save us | Ash Vadgama, AWE
Neural networks, human brains, and "The Memoristas" | Rich Vuduc, Georgia Institute of Technology
HPC and Cloud Convergence. What About HPC and Edge? | Ada Gavrilovska, Georgia Institute of Technology
How do we leverage deep learning investments to pay for HPC compiler technology? | Mary Hall, University of Utah
Brief Update: HPC/AI/QC in CN/JP/…/ | David Kahaner, Asian Technology Information Program
MPI-4: Standardization vs What Users Need for Exascale | Tony Skjellum, University of Tennessee - Chattanooga
Going data-centric in extremely heterogeneous programming - the tale of a quantum transport simulation | Torsten Hoefler, ETH Zurich
Is benchmarking worth it anymore? | Andrew Jones, NAG
Linux Must Die | Ron Minnich, Google
Experiences putting Tools into Production Codes | Olga Pearce, Lawrence Livermore National Laboratory
Do Androids Dream of Electric Sheep?, and how an Ethical AI Strategy could save us | Ash Vadgama, AWE
Neural networks, human brains, and "The Memoristas" | Rich Vuduc, Georgia Institute of Technology
HPC and Cloud Convergence. What About HPC and Edge? | Ada Gavrilovska, Georgia Institute of Technology
How do we leverage deep learning investments to pay for HPC compiler technology? | Mary Hall, University of Utah
Brief Update: HPC/AI/QC in CN/JP/…/ | David Kahaner, Asian Technology Information Program
MPI-4: Standardization vs What Users Need for Exascale | Tony Skjellum, University of Tennessee - Chattanooga
Going data-centric in extremely heterogeneous programming - the tale of a quantum transport simulation | Torsten Hoefler, ETH Zurich
Is benchmarking worth it anymore? | Andrew Jones, NAG
Linux Must Die | Ron Minnich, Google
Student Posters
Optimizing Dense Matrix Multiply for Multilevel Memory | Neil Butcher, University of Notre Dame
Adaptive Mesh Refinement in the Fast Lane | Daniel Dunning, Texas Tech University
Exploring Alternative Evaluation Methods for GPGPUs | Roland Green, Purdue University
Managing Near Memory Accelerators for Sparse Data Structures: A Software Perspective| Nelson Ho,
Performance Study of Optimization of FleCSALE Using Tabular Equation of State | Sumathi Lakshmiranganatha, University of Wyoming
Understanding Congestion in High Performance Interconnection Networks Using Sampling| Philip Taffet, Rice University
Adaptive Mesh Refinement in the Fast Lane | Daniel Dunning, Texas Tech University
Exploring Alternative Evaluation Methods for GPGPUs | Roland Green, Purdue University
Managing Near Memory Accelerators for Sparse Data Structures: A Software Perspective| Nelson Ho,
Performance Study of Optimization of FleCSALE Using Tabular Equation of State | Sumathi Lakshmiranganatha, University of Wyoming
Understanding Congestion in High Performance Interconnection Networks Using Sampling| Philip Taffet, Rice University
Session 4: Heterogeneous Applications: Maximizing Benefits While Minimizing Costs
Heterogeneous Supercomputing: HPC and the US Cancer Moonshot on Sierra | Fred Streitz, Lawrence Livermore National Laboratory
Load Balancing on Many-Core and Accelerated Systems | Cameron Smith, Rensselaer Polytechnic Institute
Exascale-Class Applications on Reconfigurable EuroEXA Prototype: Lessons Learnt and Progress So Far | Paul Carpenter, Barcelona Supercomputing Center/EuroEXA
Scientific Applications on Heterogeneous Architectures (Data Analytics and the Intersection of HPC and Edge Computing) | Michela Taufer, University of Tennessee, Knoxville
Load Balancing on Many-Core and Accelerated Systems | Cameron Smith, Rensselaer Polytechnic Institute
Exascale-Class Applications on Reconfigurable EuroEXA Prototype: Lessons Learnt and Progress So Far | Paul Carpenter, Barcelona Supercomputing Center/EuroEXA
Scientific Applications on Heterogeneous Architectures (Data Analytics and the Intersection of HPC and Edge Computing) | Michela Taufer, University of Tennessee, Knoxville
Session 5: Mapping and Optimizing Workflows for Extreme Heterogeneity
Scheduling Heterogeneous Dynamic Workloads | Tom Scogland, Lawrence Livermore National Laboratory
Workflow Management in a Heterogeneous World | Ewa Deelman, University of Southern California
Exploratory Analysis in an "In Situ World" | Janine Bennett, Sandia National Laboratories
Heterogeneous Workflows: Searching for Commonalities in the Cloud | Sam Skillman, Descartes Labs
Workflow Management in a Heterogeneous World | Ewa Deelman, University of Southern California
Exploratory Analysis in an "In Situ World" | Janine Bennett, Sandia National Laboratories
Heterogeneous Workflows: Searching for Commonalities in the Cloud | Sam Skillman, Descartes Labs