IXPUG EMEA Spring Meeting 2017

IXPUG Annual Spring Conference 2017


Location: The Stephen Hawking Centre for Theoretical Cosmology (CTC) at the University of Cambridge, UK
                 Centre for Mathematical Sciences, Wilberforce Road, Cambridge CB3 0WA, United Kingdom
 
 
Date: April 10, 2017: Intel PCC Members Only [closed session, by invite only]
April 11–13, 2017: IXPUG Annual Spring Conference Sessions [Open to public]

IXPUG is about sharing ideas, implementations, and experiences that will help users take advantage of the latest in Intel Xeon Phi technology, such as AVX512 and high-bandwidth MCDRAM memory, as well as relevant high-performance system fabrics on large-scale KNL-based systems (e.g. OmniPath).
 
 
Hosted by the The Stephen Hawking Centre for Theoretical Cosmology (CTC) at the University of Cambridge, the IXPUG Annual Spring Conference 2017 will centre on Xeon Phi-focused activity (including presentations and tutorials), where you will experience an open forum with fellow application programmers, software developers, Intel Xeon Phi architecture designers, and compiler and tool experts. Application performance and scalability challenges at all levels will be covered, including application tuning on large KNL systems.
 

Registration is now closed. Late registrations may still be accepted - for inquiries please contact Juha Jaykka <This email address is being protected from spambots. You need JavaScript enabled to view it.> directly

 

Day 1: Monday, April 10, 2017 - Intel Parallel Computing Center Members Only

Start End Title Author(s)
 12:00  13:00  Lunch/Networking  Intel Parallel Computing Center (IPCC) Members Only
 13:00  13:15  Welcome  Intel PCC Program Office
 13:15  14:45  Intel Hardware Update  Intel Staff
 14:45  15:00  Break  
 15:00  16:00  Intel Software Tools  Intel Staff
 16:00  16:30  New Intel PCCs  TBA
 16:30  17:00  Closing Remarks  Intel PCC Program Office
 

Day 2: Tuesday, April 11, 2017 - IXPUG Annual Spring Conference 

Start End Title Author(s) Presentation Video
 8:30  9:00  Registration & Coffee/Tea      
 9:00  9:10  Welcome from the IXPUG Spring 2017 Chair  Professor Paul Shellard,The Stephen Hawking Centre for  Theoretical Cosmology at University of Cambridge   PDF  
 9:10  9:20  Welcome from the IXPUG Chair  Michael Lysaght, ICHEC  PDF  
 9:20  10:00  Keynote: Parallelism, why and how?  James Cownie, Intel Corporation   PDF  Video
 10:00  10:30  Site Updates (15 minutes each)

 - MARCONI, Fabio Affinito, CINECA

 - Cambridge, Gilles Civario, DellEMC

 PDF  Video
 10:30  11:00  Early Science with KNL  Richard Gerber, NERSC  PDF  Video
 11:00  11:30  Tea Break      
           
 11:30  12:45  IXPUG Lightning Talks (15 minutes each)
  • LT1: Scaling Distributed Neural Network Training on Large HPC Clusters, Valeriu Codreanu, Damian Podareanu and Vikram Saletore
  • LT2: Characterizing Deep Learning Training Workloads on Intel Knights Landing - Using TensorFlow and Caffe Case Studies, Gaurav Kaul, Jimmy Tran and Ryo Asai
  • LT3: Benchmarking performances of the Xeon-Phi KNL with a CFD code, Alessandro Gabbana
  • LT4: Optimizing Elmer finite element software on KNL, Juhani Kataja, Mikko Byckling and Thomas Zwinger

 LT1 PDF

 LT2 PDF

 LT3 PDF

 LT4 PDF

 Video

 Video

 Video

 Video

 12:45  14:00  Lunch      
 14:00  14:30  IXPUG Presentation 1  Pre-exascale architectures: Many-core  Intel KNL  performance and usability  assessment for French  scientific  communities, Gabriel Hautreux and Eric  Boyer  PDF  Video
 14:30  15:00  IXPUG Presentation 2  Performance optimization of  OpenFOAM on KNL,  Ravi  Ojha, Paul  Edwards, Prasad Pawar, Nishant  Agrawal  and  Sonia Rani  PDF  Video
 15:00  15:30  IXPUG Presentation 3  SeisSol on KNL, Carsten Uphoff and  Michael Bader   PDF  Video
 15:30  16:00  IXPUG Presentation 4  Grid: Structured Cartesian Mesh  Library for  Quantum  Chromodynamics, Peter Boyle, Guido  Cossu, Antonin  Portelli and Azusa  Yamaguchi   PDF  Video
 16:00  16:30  Tea Break       
 16:30   17:00  IXPUG Presentation 5  Deep Learning under the Hood,  Sidharth N. Kashyap,  Angus Lepper, Peter  Boyle  PDF  Video
           
 17:00  18:00  Lightning Talks (15 minutes each)
  • LT6: An efficient SIMD implementation of pseudo Verlet-lists for neighbour interactions in particle-based codes, James Willis and Matthieu Schaller
  • LT7: Special functions in quantum chemistry, Vladimir Mironov and Alexander Moskovsky
  • LT8: Numerical general relativity on Intel KNL, Alejandro Duran, Juha Jäykkä, Markus Kunesch and Saran Tunyasuvunakool

 LT6 PDF

 LT7 PDF

 LT8 PDF

 Video

 Video

 Video

 18:15  19:45  Networking Reception      
 
 

Day 3: Wednesday, April 12 2017 - IXPUG Annual Spring Conference

Start End Title Author(s) Presentation Video
 8:30  9:00  Registration & Coffee/Tea      
 9:00  9:30  IXPUG Presentation 6  Optimizing Vlasiator on KNL, Sebastian  von Alfthan   PDF  Video
 9:30  10:00  IXPUG Presentation 7  The GeantV prototype on KNL, Andrei  Gheata, Sofia  Vallecorsa and Federico  Carminati   PDF  Video
 10:00  10:30  IXPUG Presentation 8  BEM quadrature schemes for many-  core  architectures, Jan  Zapletal, Michal  Merta and Lukas  Maly  PDF  Video
 10:30  11:00 IXPUG Working Groups 2017 & Discussion  Board  Michael Lysaght, John Pennycook  PDF  
 11:00  11:30  Tea Break       
 11:30  12:00  IXPUG Presentation 9  Task-based parallelism in  DL_POLY_4, Aidan Chalk,  Alin  Marin  Elena and Luke Mason   PDF  Video
 12:00  13:00  IXPUG Lightning Talks (15 minutes each)
  • LT9: Deeper Insight with Intel Advisor’s Cache Aware Roofline Model, Michael Lysaght
  • LT10: Running containers on Intel® Xeon Phi™ processors, Piotr Uminski and Jakub Dlugolecki
  • LT11: Ensuring program correctness and optimal performance on Intel Knights Landing with Allinea tools, Patrick Wohlschlegel

 LT9 PDF

 LT10 PDF

 LT11 PDF

 Video

 Video

 Video

 13:00  14:00  Lunch Break      
 14:00  14:30  IXPUG Presentation 10  The MEGA-STREAM benchmark on  Intel Xeon Phi  processors (Knights  Landing), Tom Deakin, John  Pennycook,  Andrew Mallinson, Wayne Gaudin and  Simon  Mcintosh-Smith   PDF  Video
 14:30  15:00  IXPUG Presentation 11  Vectorization and locality  optimization for seismic  imaging  through automatic code generation,  Navjot  Kukreja, Fabio Luporini, Mathias  Louboutin, Michael  Lange  and  Gerard  Gorman  PDF  Video
 15:00  15:30  IXPUG Presentation 12  Improving achieved memory  bandwidth from C++  codes  on Intel®  Xeon Phi™ Processor (Knights  Landing), Karthik  Raman, Tom Deakin,  James Price  and Simon Mcintosh-Smith  PDF  Video
 15:30  16:00  IXPUG Presentation 13  Acceleration of Blender Cycles Render  Engine using  Intel  Xeon Phi , Milan  Jaros and Michal Merta  PDF  Video
 16:00  16:30  Tea Break      
           
 16:30 17:30  IXPUG Lightning Talks (15   minutes each)
  • LT12: High Fidelity and High Performance Visualization, Johannes Günther
  • LT13: Simultaneous offload using multiple accelerators, Matthias Noack and Kacper Kornet
  • LT14: Offloading over Fabric for the Timothy application: Running agent-based parallel tissue modeling framework on a heterogeneous system with Intel® Xeon® and Xeon Phi™ Processors, Piotr Uminski, Maciej Cytowski, Zuznna Szymanska, Krzysztof Raszkowski and Grzegorz Andrejczuk

 LT12 PDF

 LT13 PDF

 LT14 PDF

 Video 

 Video

 Video

 17:30  18:00  Panel Discussion      
 18:00  18:10  Closing Remarks  Michael Lysaght, ICHEC    
 19:00    Conference Dinner - TBD      
 
 

Day 4: Thursday April 13, 2017: IXPUG Annual Spring Conference

Start End Title Author(s) Presentation Video
 8:30  9:00  Registration & refreshments      
 9:00  12:30  Tutorial I  Intel  PDF  
 12:30  14:00  Lunch Break      
 14:00  17:00  Tutorial II  Intel  PDF  
 
 
 
 
Important Dates:
Abstract Submission Deadline 26 February 2017 (AoE), extended to 5 March 2017
Abstracts Reviewed by IXPUG Committee 27 Feb - 3 March 2017, extended to 5 March - 10 March 2017
Acceptance Notification 7 March 2017, extended to 14 March 2017
Preliminary Agenda Posted to IXPUG Website 10 March 2017, extended to 17 March 2017
Registration Deadline 24 March 2017
Agenda Finalized (Presenter Notifications) 31 March 2017
Final Presentations Due from Speakers 5 April 2017
Intel PCC Sessions 10 April 2017
IXPUG Sessions 11-13 April 2017
 
 

Travel Information:

  • Visa Requirements: Please contact This email address is being protected from spambots. You need JavaScript enabled to view it. at University of Cambridge for visa inquiries.
  • Transportation, Parking, etc.: Information on how to get to the venue can be found here.
  • Lodging: University rooms can be booked here and additional hotel recommendations listed below. A useful site for finding accommodation here
Hotel Hotel Address

Arundel House Hotel

Chesterton Road, Cambridge CB4 3AN

Phone: +44 (0) 1223 367 701
Varsity Hotel

Thompson’s Lane (off Bridge Street) Cambridge CB5 8AQ

Phone: +44 (0) 1223 30 60 30

Hotel Felix

Whitehouse Lane, Huntingdon Road,Cambridge, CB3 0LX

Phone: +44 01223 277977
Hilton (City Centre)

Downing Street, Cambridge, CB2 3DT, United Kingdom

Phone: +44 1223 464491
Hilton (By the river)

Granta Place Mill Lane, Cambridge,CB2 1RT, United Kingdom

Phone: +44 1223 259988

  

Organizing Committee and Contacts

Organizing Committee:

  • Juha Jaykka, This email address is being protected from spambots. You need JavaScript enabled to view it., Department of Applied Mathematics and Theoretical Physics, University of Cambridge, UK
  • Estela Suarez, This email address is being protected from spambots. You need JavaScript enabled to view it., Forschungszentrum Juelich / Juelich Supercomputing Centre, Germany
  • Thomas Steinke, This email address is being protected from spambots. You need JavaScript enabled to view it., Zuse Institute Berlin, Germany
  • David Martin, This email address is being protected from spambots. You need JavaScript enabled to view it., Argonne National Laboratory, US
  • John Pennycook, This email address is being protected from spambots. You need JavaScript enabled to view it., Intel Corporation, United States
  • Richard Gerber, This email address is being protected from spambots. You need JavaScript enabled to view it.National Energy Research Scientific Computing Center, Lawrence Berkeley National Lab
  • Michael Lysaght, This email address is being protected from spambots. You need JavaScript enabled to view it. , Irish Centre for High End Computing (ICHEC), Ireland

Local Organizer:

  • James Parke, This email address is being protected from spambots. You need JavaScript enabled to view it., University of Cambridge
 

Call for Presentations:
IXPUG welcomes submissions on innovative work from KNL users in academia, industry and government labs, describing original discoveries and experiences that will promote and prescribe efficient use of many-core and multicore systems. The authors of the best scored abstracts and draft presentations will be selected for a full 30 minute presentation; others may be offered an opportunity to present shorter Lightning Talks.

 

Submission Guidelines:
A short Abstract and Draft Presentation should be submitted by Sunday, 26th February. Sunday, March 5 [Deadline extended 1 week!] The Draft Presentation does not need to be complete by this date. However, along with the Abstract, it should reflect the overall intent of the presentation and contain placeholders for the remaining content to be completed by the Final Presentation Deadline on 5th April. Presentations describing application results on multi-node configurations/addressing KNL-specific features (e.g. use of MCDRAM)/peformance tools exploitation will be prioritized. For presentation format, please use the IXPUG presentation template and you can submit your abstract to IXPUG EasyChair.

 

Topics of interest are (but not limited to):

  • Vectorization: Data layout for efficient SIMD operations, SIMD directives and operations
  • Memory: Data layout in memory for efficient access and caching (data preconditioning), access latency concerns (prefetch, streams, costs for MCDRAM), partitioning of DDR and MCDRAM for applications (memory policies)
  • Communication: including early experiences with Omni-Path
  • Thread and Process Management: Process and thread affinity issues, SMT (simultaneous multi-threading, in core), balancing processes and threads, nested threading (per tile or core)
  • Multi-node Application Experience: especially on large-scale KNL systems
  • Programming Models: OpenMP 4.x, hStreams, MPI 3.0, hybrid programming (e,g, MPI/OpenMP,), others
  • Algorithms and Methods: including scalable and vectorizable algorithms
  • Software Environments and Tools
  • Benchmarking & Profiling Tools
  • Visualization

 

Preliminary Agenda Schedule (final agenda will be published after abstract submission review):

  • Monday, April 10, 2017: Intel Parallel Computing Center (Intel PCC) members only meeting, light lunch during 12:00 - 13:30, followed by Intel PCC sessions (13:30-1700)
  • Tuesday, April 11, 2017: IXPUG Annual Spring Conference, keynote, presentations, etc.
  • Wednesday, April 12, 2017: IXPUG Annual Spring Conference, presentations, followed by conference dinner
  • Thursday, April 13, 2017: IXPUG Annual Spring Conference, tutorial day