CUG2018 Proceedings


Overview | By Event Type | Author Index

Monday, May 21st


8:30am-10:00am

Tutorial 1B
Event Small
Harold Longley
Managing SMW/eLogin and ARM XC nodes
Tutorial

Tutorial 1C
Studio 6
Michael Ringenburg
Analytics and Artificial Intelligence Workloads on Cray Systems
Tutorial

Tutorial 1D
Studio 3
John Levesque
Applying a “Whack-a-Mole” Method Using Cray perftools to Identify the Moles
Tutorial

10:30am-12:00pm

Tutorial 1B Continued
Event Small
Harold Longley
Managing SMW/eLogin and ARM XC nodes
Tutorial

Tutorial 1C Continued
Studio 6
Michael Ringenburg
Analytics and Artificial Intelligence Workloads on Cray Systems
Tutorial

Tutorial 1D Continued
Studio 3
John Levesque
Applying a “Whack-a-Mole” Method Using Cray perftools to Identify the Moles
Tutorial

1:00pm-2:30pm

Tutorial 2B
Event Small
Harold Longley
Managing SMW/eLogin and ARM XC nodes
Tutorial

Tutorial 2C
Studio 6
Michael Ringenburg
Analytics and Artificial Intelligence Workloads on Cray Systems
Tutorial

Tutorial 2D
Studio 3
Benjamin Landsteiner
DataWarp Administration Tutorial
Tutorial

3:00pm-4:30pm

Tutorial 2B Continued
Event Small
Managing SMW/eLogin and ARM XC nodes
Tutorial

Tutorial 2C Continued
Studio 6
Michael Ringenburg
Analytics and Artificial Intelligence Workloads on Cray Systems
Tutorial

Tutorial 2D Continued
Studio 3
DataWarp Administration Tutorial
Tutorial

4:40pm-6:40pm

BoF 3A
Event Large
Bilel Hadri
Programming Environments, Applications, and Documentation (PEAD) Special Interest Group meeting
Birds of a Feather

BoF 3B
Event Small
Nicholas Cardo
Systems Support SIG Meeting
Birds of a Feather

BoF 3C
Studio 6
Michele Bertasi; Torben Kling Petersen
When to use Flash and when not to …..
Scalable Accounting & Reporting for Compute Jobs
Birds of a Feather

BoF 3D
Studio 3
Sadaf R. Alam
Tools and Utilities for Data Science Workloads and Workflows
Opportunities for containers in HPC ecosystems
Birds of a Feather

Tuesday, May 22nd


9:40am-9:50am

Sponsor Talk 5
Event Large
Trey Breckenridge
Software ecosystem for Arm-based HPC
Sponsor Talk

9:50am-10:00am

Sponsor Talk 6
Event Large
Trey Breckenridge
PBS Works 2018 and Beyond: Multi-sched, Power Rate Ramp Limiting, Soft walltime, Job Equivalence Classes, …, Native Mode
Sponsor Talk

1:00pm-2:30pm

Technical Session 8A
Event Large
Ronald Brightwell
Cray Next Generation Software Integration Options
CLE Port to ARM: Functionality, Performance, and Lessons Learned
SSA, ClusterStor Call-home Service Actions, and an Introduction to Cray Central Telemetry and Triage Services (C2TS)
Paper

Technical Session 8B
Event Small
Bilel Hadri
Performance Evaluation of MPI on Cray XC40 Xeon Phi Systems
Performance Impact of Rank Reordering on Advanced Polar Decomposition Algorithms
Are We Witnessing the Spectre of an HPC Meltdown?
Paper

Technical Session 8C
Studio 6
Frank M. Indiviglio
How Deep is Your I/O? Toward Practical Large-Scale I/O Optimization via Machine Learning Methods
TOKIO on ClusterStor: Connecting Standard Tools to Enable Holistic I/O Performance Analysis
Improving Nektar++ IO Performance for Cray XC Architecture
Paper

3:00pm-5:00pm

Technical Session 9A
Event Large
Ann Gentile; Jim Rogers
Cray System Monitoring: Successes, Priorities, Visions
Use of the ERD for administrative monitoring of Theta
Supporting failure analysis with discoverable, annotated log datasets
Paper

Technical Session 9B
Event Small
Zhengji Zhao
Intel® Xeon processor Scalable family performance in HPC, AI and other key segments
Storage and memory hierarchy in HPC: new paradigm and new solutions with Intel
Chapel Comes of Age: Making Scalable Programming Productive
The Cray Programming Environment: Current Status and Future Directions
Paper

Technical Session 9C
Studio 6
Ronald Brightwell
Scaling Deep Learning without Impacting Batchsize
Alchemist: An Apache Spark <=> MPI Interface
Continuous integration in a Cray multiuser environment
Paper

5:10pm-6:20pm

BoF 10A
Event Large
Larry Kaplan
Cray Next Generation Software Integration Options
Birds of a Feather

BoF 10B
Event Small
Bilel Hadri
Managing Effectively the User Software Ecosystem
Birds of a Feather

BoF 10C
Studio 6
Stephen Leak
Practical implementation of monitoring on Cray system
Birds of a Feather

BoF 10D
Studio 3
David Hancock
Open Discussion with CUG Board
Birds of a Feather

Wednesday, May 23rd


9:40am-9:50am

Sponsor Talk 12
Event Large
Trey Breckenridge
PGI Compilers for Accelerated Computing
Sponsor Talk

9:50am-10:00am

Sponsor Talk 13
Event Large
Trey Breckenridge
Making a Large Investment in HPC? Think Bright
Sponsor Talk

11:50am-12:00pm

New Site 15
Event Large
Helen He
Isambard, the world’s first production Arm-based supercomputer
New Site

2:00pm-2:10pm

New Site 17
Event Large
Helen He
The NIWA/NeSI HPC Replacement Project - A Voyage into Complexity: Integrating multi-site XC, CS, ESS, Spectrum Scale, and OpenStack
New Site

2:10pm-2:20pm

Sponsor Talk 18
Event Large
Trey Breckenridge
Applying DDN to Machine Learning
Sponsor Talk

2:20pm-2:30pm

Sponsor Talk 19
Event Large
Trey Breckenridge
Slurm Overview and Road Map
Sponsor Talk

3:00pm-5:00pm

Technical Session 20A
Event Large
Jim Rogers
Modernizing Cray Systems Management – Use of Redfish APIs on Next Generation Shasta Hardware
Managing the SMW as a git Branch
External Login Nodes at Scale
Best Practices for Management and Operation of Large HPC Installations
Paper

Technical Session 20B
Event Small
Zhengji Zhao
Comparative Benchmarking of the First Generation of HPC-Optimised Arm Processors on Isambard
Evaluating Runtime and Power Requirements of Multilevel Checkpointing MPI Applications on Four Different Parallel Architectures: An Empirical Study
OpenACC and CUDA Unified Memory
Strategies to Accelerate VASP with GPUs Using OpenACC
Paper

Technical Session 20C
Studio 6
Jim Williams
Cray Storage Road Map
DataWarp Transparent Cache: Implementation, Challenges, and Early Experience
PBS Professional - Optimizing the "When & Where" of Scheduling Cray DataWarp Jobs
DataWarp Transparent Caching: Data Path Implementation
Paper

5:10pm-6:20pm

BoF 21A
Event Large
Harold Longley
XC System Management Usability BOF
Birds of a Feather

BoF 21B
Event Small
Chris Fuson
Best Practices for Supporting Diverse HPC User Storage Needs with Finite Storage Resources
Birds of a Feather

BoF 21C
Studio 6
Michael Showerman
Automated Analysis and Effective Feedback
Birds of a Feather

BoF 21D
Studio 3
Kelly J. Marquardt
Customer Collaboration and the Shasta Software Stack
Birds of a Feather

Thursday, May 24th


9:50am-10:00am

Sponsor Talk 23
Event Large
Trey Breckenridge
AMD update - EPYC !
Sponsor Talk

10:30am-12:00pm

Technical Session 24A
Event Large
Jim Rogers
Trinity: Opportunities and Challenges of a Heterogeneous System
Cray Advanced Power Management Updates
Weathering the Storm – Lessons Learnt in Managing a 24x7x365 HPC Delivery Platform
Paper

Technical Session 24B
Event Small
David Paul
Using CAASCADE and CrayPAT for Analysis of HPC Applications
Toward Automated Application Profiling on Cray Systems
Roofline Analysis with Cray Performance Analysis Tools (CrayPat) and Roofline-based Performance Projections for a Future Architecture
Paper

Technical Session 24C
Studio 6
Tim Robinson
Enabling Docker for HPC
Installation, Configuration and Performance Tuning of Shifter V16 on Blue Waters
Incorporating a Test and Development System Within the Production System
Paper

1:00pm-2:30pm

Technical Session 25A
Event Large
Tim Robinson
TensorFlow at Scale - MPI, RDMA and All That
High Performance Scalable Deep Learning with the Cray Programming Environments Deep Learning Plugin
Performance evaluation of parallel computing and Big Data processing with Java and PCJ library
Paper

Technical Session 25B
Event Small
Frank M. Indiviglio
Leveraging MPI RMA to optimise halo-swapping communications in MONC on Cray machines
Cray Performance Tools for Analyzing Applications at Scale
Performance Study of Popular Computational Chemistry Softwares on Cray HPC
Paper

Technical Session 25C
Studio 6
Paul L. Peltz Jr.
The Role of SSD Block Caches in a World with Networked Burst Buffers
Use of View for ClusterStor to monitor and diagnose performance of your Lustre filesystem
Nuclear Meltdown?: Assessing the impact of the Meltdown/Spectre bug at Los Alamos National Laboratory
Paper

3:00pm-4:30pm

Technical Session 26A
Event Large
Bilel Hadri
Optimised all-to-all Communication on Multicore Architectures Applied to FFTs with Pencil Decomposition
Eigensolver Performance Comparison on Cray XC Systems
On the Use of Vectorization in Production Engineering Workloads
Paper

Technical Session 26B
Event Small
Jim Williams
Unikernels/Library Operating Systems for HPC
GPU Usage Reporting
Instrumenting Slurm Command Line Commands to Gain Workload Insight
Paper

Technical Session 26C
Studio 6
Chris Fuson
How to implement the Sonexion RestAPI and correlate it with SEDC and other data.
Usage and Performance of libhio on XC-40 Systems
Improved I/O Using Native Spectrum Scale (GPFS) Clients on a Cray XC System
Paper

Created 2018-9-1 19:3