IXPUG Europe Spring 2018



IXPUG Annual Spring Conference 2018

Date: March 5-7, 2018

Location: CINECA head office in Bologna, Italy (via Magnanelli, 6/3 - 40033 Casalecchio di Reno (Bologna))

Registration: Registration is now closed.


Event Description: The Intel eXtreme Performance Users Group (IXPUG) is an active community led forum for sharing industry best practices, techniques, tools, etc. for maximizing efficiency on Intel platforms and products. This meeting will focus on all aspects of employing and adopting manycore processing technologies and techniques for optimal application execution.


Travel Details:

·         Visa Requirements: Please contact This email address is being protected from spambots. You need JavaScript enabled to view it. for visa inquiries. U.S. citizens may enter Italy for up to 90 days for tourist or business purposes without a visa.

·         Public Transportation: visit the CINECA website for public transportation recommendations.

·         Hotel Recommendations: nearby hotels suggestions can be found HERE

 Watch remotely: the IXPUG meeting will be live streamed at this URL: https://streaming.cineca.it/DefaultPlayer/div.php?evento=IXPUG2018 


Event Agenda

Day 1: Monday, March 5, 2018

Start End Title Author(s) Presentation
09:00 09:30am Registration & Coffee/Tea    
09:30 09:45 Welcome by Host Fabio Affinito  PDF
09:45 10:00am Welcome by IXPUG Board Lisa M. Smith  PDF
10:00am 10:45am

Keynote Presentation:

Electronic Structure Codes on KNL: Structural, electronic and optical properties of graphene Nanoribbons

Andrea Ferretti - CNR S3 Center


10:45 11:15am

Site Updates: TACC, Argonne and Juelich

John Cazes




11:15 11:30am


11:30 12:00pm

Impact of Meltdown/Spectre Patches on HPC Application Performance

John Cazes, Lei Huang and Kevin Chen PDF
12:00 12:30pm

Electronic structure codes on KNL, profiling and benchmarking results

Pietro Bonfa, Fabio Affinito and Carlo Cavazzoni PDF
12:30 13:30


13:30 14:00

GPAW Performance Optimisation and Energy Consumption on KNLs

Martti Louhivuori PDF
14:00 14:30

Large-Scale Implementation of the Density Matrix Renormalization Group Algorithm: Performance Analysis on the Intel Broadwell and Knights Landing Architectures

James Vance, Ivan Girotto and Marcello Dalmonte PDF
14:30 14:45

Porting of the DBCSR Library for Sparse Matrix-Matrix Multiplications to Intel Xeon Phi Systems

Alfio Lazzaro and Juerg Hutter

14:45 15:15


Tour the CINECA facility (Andrew/Fabio)

*Sign up at registration desk 

15:15 15:45

LIBXSMM: Look Inside at Top-Notch Low-Level Software Primitives

Hans Pabst PDF
15:45 16:15

A Performance Comparison of Deep Learning Frameworks on KNL

Riccardo Zanella, Giuseppe Flameni and Marco Rorro PDF
16:15 16:45

Machine Learning for Fast Detector Simulation


Sofia Vallecorsa, Federico Carminati, Gulrukh Khattak and Andrea Zanetti

16:45 17:00

IXPUG Day 1 Closing Remarks

(Recap of Day 1, Agenda for Day 2)

Fabio Affinito  

Networking Reception

Cafe Zanarini

(Bus from CINECA to be provided)


Day 2: Tuesday, March 6, 2018 

Start End Title Author(s) Presentation
09:00am 09:30  Registration & Coffee/Tea    
09:30 09:45am  Day 2 Agenda review  Fabio Affinito  
09:45 10:15

Porting, Optimization and Bottlenecks of OpenFOAM in KNL Environment 

Ivan Spisso, Giorgio Amati, Vittorio Ruggiero and Carlo Fiorina

10:15 10:45 Devito - Automated High-Performance Finite-Difference for Geophysical Exploration Fabio Luporini, Charles Yount, Michael Lange, Navjot Kukreja, Mathias Louboutin, Philipp Witte, Gerard Gorman and Felix Herrmann PDF
10:45 11:15 OpenMP Threading and Vectorization of MPI Finite Element Code Elmer Mikko Byckling and Juhani Kataja PDF
11:15 11:30am Break

CINECA Facility tour #2 (Andrew/Fabio)

*sign up at registration 

11:30am 12:00pm Optimization of D3Q19 Lattice Boltzmann Kernels for Recent Multi- and Many-cores Intel Based Systems Ivan Girotto, Sebastiano Fabio Schifano and Federico Toschi PDF
12:00 12:30pm Lunch    
12:30 13:00 Updates on Software-Defined Visualization Johannes Günther PDF
13:00 13:30 Software-Defined Visualization in MegaMol Tobias Rau, Patrick Grakla and Guido Reina PDF
13:30 14:00 Rendering in Blender Cycles using AVX512 Vectorization Milan Jaroš, Petr Strakoš and Lubomir Říha  PDF
14:00 14:15 Break
Group photo!
14:15 14:45 Optimization of the Gadget Code and Energy Measurements on Second-Generation Intel Xeon Phi Luigi Iapichino  PDF
14:45 14:45

Panel Discussion: Putting It All Together: Scientific Workflows at Scale 

Panelists: Heinrich Bockhorst, Klaus-Dieter Oertel, Johannes Günther and Hans Pabst  
14:45 15:00

IXPUG Closing Remarks 

(Day 2 Recap and Day 3 Agenda)

Fabio Affinito  


Day 3: Wednesday, March 7, 2018 

Start End Title Author(s) Presentation
 08:30am 09:00 Registration and Coffee/Tea    
09:00 09:15 Day 3 Agenda Review Fabio Affinito  
09:15am 10:45 Tutorial 1: Intel Application Snapshot and Intel VTune Amplifier Tutorial Heinrich Bockhorst and Klaus-Dieter Oertel PDF
10:45 11:00am Break    
11:00am 12:30pm Tutorial 2: Intel Vector Advisor  Heinrich Bockhorst and Klaus-Dieter Oertel PDF
12:30 13:00 Lunch    
13:00 15:00 Open Labs Heinrich Bockhorst, Klaus-Dieter Oertel and Hans Pabst  


 Submissions are now closed. The agenda will be updated following the notification of presenters.


Important dates: 

Abstract submissions open: January 3, 2018

Abstracts due: January 31, 2018  February 2, 2018 (AoE) 

Presenters notified: February 7, 2018  February 14, 2018 (AoE)

Final presentations: March 2, 2018 


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 technical lecture; others may be offered an opportunity to present shorter 10-15 minute Lightning Talks.


Submission Guidelines:  Submissions are now closed. The abstract submission process opened on January 3, 2018 and will close on January 31, 2018. All submitters must upload presentations to the IXPUG EasyChair site. The final presentations are due on March 2, 2018 and the content should reflect the overall intent outlined in the abstract submission. Please be sure to focus your content on the approach that was taken, obstacles encountered, solutions developed and ultimate results. Feel free to include application results on multi-node configurations addressing KNL-specific features (e.g. use of MCDRAM)/performance tools exploitation will be prioritized.  You may wish to refer to presentations from previous events for appropriate elements.


Topics of interest are (but not limited to): sharing techniques in vectorization, memory, communications, thread and process management, multi-node application experiences, programming models, algorithms and methods, software environment and tools, benchmarking and profiling tools, visualization development, etc.