CERN openlab Technical Workshop

Europe/Zurich
503/1-001 - Council Chamber (CERN)

503/1-001 - Council Chamber

CERN

31
Show room on map
Maria Girone (CERN)
Description

During this event — which will take place over two days — we will review the CERN openlab projects carried out in the first year of our sixth three-year phase, with technical discussion of recent progress made. 

Talks will be organised into sessions focused on the four R&D topics set out in CERN openlab's recent white paper: (1) data centre technologies and infrastructures, (2) computing performance and software, (3) machine learning and data analytics, (4) applications in disciplines beyond high-energy physics. There will be an additional fifth session of talks at the workshop, centred on the topic of quantum computing.

The latest technical posters from the CERN openlab projects will also be presented at the event.

Participants
  • Ahmad Hesam
  • Aimilios Tsouvelekakis
  • Alberto Aimar
  • Alessandro Raimondo
  • Alexey Fishkin
  • Amy Bilton
  • Andrea Manzi
  • Andreas Alexopoulos
  • Andrew Robert Purcell
  • Anna Zaborowska
  • Anthony Hennessey
  • Antonio Nappi
  • Ben Segal
  • Benedikt Würkner
  • Benjamin Frisch
  • BİRANT BARAN
  • Brice Copy
  • Claudio Bellini
  • Concezio Bozzi
  • Cosimo Damiano Gianfreda
  • Cris Pedregal-Martin
  • Dan van der Ster
  • Daniel Hugo Campora Perez
  • Daniele Gregori
  • Danilo Cicalese
  • Dario d'Andrea
  • David Ebert
  • David Sangcheol Lee
  • David Widegren
  • Davide Costanzo
  • Douglas Gibbs
  • Edoardo Martelli
  • Elena Graverini
  • Eleonora Villani
  • Emiliano Piselli
  • Enrico Gamberini
  • Eric Aquaronne
  • Eric Grancher
  • Eugenio Culurciello
  • Eva Dafonte Perez
  • Fabio Luchetti
  • Federico Carminati
  • Felice Pantaleo
  • Filippo tilaro
  • Fons Rademakers
  • Francesco Fiori
  • Francisco Perez
  • Frank Berghaus
  • Gabrielle Hugo
  • Gancho Dimitrov
  • Georgios Kaklamanos
  • Germano Massullo
  • Graeme Stewart
  • Gregor Molan
  • Gurjeet Singh
  • H. Peter Hofstee
  • Helge Meinhard
  • Hok Chuen Cheng
  • Ingo Thon
  • Jacopo Pazzini
  • Jacopo Pinzino
  • Jakob Blomer
  • Jason Adlard
  • Johannes Frank
  • Jose Carlos Luna Duran
  • Josef Spillner
  • Josep Soler Garrido
  • José Márquez
  • Julien Collet
  • Karim Massri
  • Kristina Gunne
  • Kyle Poland
  • Lance Thompson
  • Lionel Clavien
  • Lorenzo Moneta
  • Luca Atzori
  • Luca Canali
  • Luca Mascetti
  • Luis Rodriguez Fernandez
  • Lukas Felsberger
  • Maite Barroso Lopez
  • Manuel Rodriguez
  • Marco Zanetti
  • Maria Girone
  • Marie-Christine Sawley
  • Mario Lassnig
  • Mark Gray
  • Mark Hur
  • Markus Schulz
  • Martin Barisits
  • Mathieu Saccani
  • Matteo Lupi
  • Matteo Migliorini
  • Mattia Fani'
  • Maurizio Pierini
  • Michel Roethlisberger
  • Monika .
  • Monika Grothe
  • Nikola Hardi
  • Nikolaos Androulakis
  • Nuri Twebti
  • Olga Vladimirovna Datskova
  • Pablo Llopis Sanmillan
  • Panagiotis Gkikopoulos
  • Patrick Duverger
  • Peter Elmer
  • Peter Szegedi
  • Piero Altoe
  • Pieter Van Trappen
  • Piotr Golonka
  • Remi Mommsen
  • Ricardo Picatoste Ruilope
  • Riccardo Castellotti
  • Richard Hay Jr.
  • Rosen Matev
  • Ruben Domingo Gaspar Aparicio
  • samuele altruda
  • Sebastian Lopienski
  • Servesh Muralidharan
  • Simon Eitelbuss
  • Sonia BEN HAMIDA
  • Srdjan Atanasijevic
  • Taghi Aliyev
  • Tatiana Mangels
  • Tibor Simko
  • Tim Bell
  • Tim Leonhardt
  • Tobias Schuele
  • Tom Gibbs
  • Tristan Roberge-Mentec
  • Ugo Gentile
  • Ugo Varetto
  • Utsav Siwakoti
  • Vaggelis Motesnitsalis
  • Viktor Kovacevic
  • vincent leocorbo
  • Vincenzo Innocente
  • Wainer Vandelli
  • Wassef Karimeh
  • Yunus Emre Alpu
  • Wednesday, 23 January
    • 09:30 09:45
      Welcome and goals of the workshop 15m 503/1-001 - Council Chamber

      503/1-001 - Council Chamber

      CERN

      31
      Show room on map
      Speaker: Maria Girone (CERN)
    • 09:45 10:45
      Computing architectures for machine learning, data acquisition and processing part 1 503/1-001 - Council Chamber

      503/1-001 - Council Chamber

      CERN

      31
      Show room on map
    • 10:45 11:15
      Coffee 30m 503/1-001 - Council Chamber

      503/1-001 - Council Chamber

      CERN

      31
      Show room on map
    • 11:15 12:15
      Computing architectures for machine learning, data acquisition and processing part 1 503/1-001 - Council Chamber

      503/1-001 - Council Chamber

      CERN

      31
      Show room on map
      • 11:15
        Partner talk Siemens: AI on the machine level in industrial automation 20m

        Artificial intelligence, with all its different facets, makes a considerable contribution, especially in industry, toward reducing the usual expense of programming and engineering, making the control logic more agile and flexible with regard to changes in the ambient conditions and structuring production processes with greater flexibility and precision. With Future of Automation, Siemens is offering far-reaching insights into the future of automation and the role of artificial intelligence within the portfolio of Totally Integrated Automation. This means scalable solutions from the field level to the controller and edge level and all the way to the Cloud. This means that an AI solution can be scaled in terms of the environment and the target application: At the machine on the field level where fast, deterministic decisions are required, or across all machines or plants with a significantly higher quantity of data to be processed and a corresponding demand for computing power. To enable AI at the lowest level Siemens introduced a Technology module, the S7-1500 TM NPU (neural processing unit), which enables the efficient processing of neural networks. This allows to use machine-learning algorithms, for example, visual quality checks in production plants or image-guided robot system. This allows a considerably more efficient and more "human-like" behaviour possible.

        Speakers: Ingo Thon (Siemens) , Jose Soler Garrido (Siemens)
      • 11:35
        Intel fast simulation 20m
        Speaker: Federico Carminati (CERN)
      • 11:55
        IBM Evaluation of power architectures for machine learning 20m
        Speakers: Ahmad Siar Hesam (Technische Universiteit Delft (NL)) , Daniel Hugo Campora Perez (Universidad de Sevilla (ES))
    • 12:15 13:30
      Lunch and poster session 1h 15m 503/1-001 - Council Chamber

      503/1-001 - Council Chamber

      CERN

      31
      Show room on map
    • 13:30 15:10
      Computing architectures for machine learning, data acquisition and processing part 2 503/1-001 - Council Chamber

      503/1-001 - Council Chamber

      CERN

      31
      Show room on map
      • 13:30
        Micron: Exploring Accelerated Machine Learning for Experiment Data Analytics 20m
        Speakers: Emilio Meschi (CERN) , Maurizio Pierini (CERN)
      • 13:50
        Partner talk Micron: Advanced Computing Solutions 20m

        Micron introduces machine learning products designed to accelerate deep learning algorithm in hardware.
        Coupled with FWDNXT Inference Engine, the programmable logic products from Micron offer the ability to execute complex neural networks in hardware.
        We present expertise in neural network design, compiler and hardware acceleration. We present FWDNXT SDK software as an alternative to graphic processors for deep learning acceleration.

        Speaker: Mark Hur (Micron)
      • 14:10
        E4: A Testbed for GPU Accelerated Applications 20m
        Speaker: Felice Pantaleo (CERN)
      • 14:30
        Status of the DEEP-EST project and outlook 20m
        Speaker: Viktor Khristenko (CERN)
      • 14:50
        Intel: Fast deep neural network inference on FPGAs 20m
        Speaker: Jennifer Ngadiuba (CERN)
    • 15:10 15:50
      Coffee 40m Pas Perdus

      Pas Perdus

    • 15:50 17:35
      Data Center Technologies 503/1-001 - Council Chamber

      503/1-001 - Council Chamber

      CERN

      31
      Show room on map
      • 15:50
      • 16:10
        Comtrade EOS productization 20m
        Speaker: Luca Mascetti (CERN)
      • 16:30
        Oracle Management Cloud: A unified monitoring platform 20m
        Speaker: Aimilios Tsouvelekakis (Ministere des affaires etrangeres et europeennes (FR))
      • 16:50
        Running JAVA application servers on Kubernetes 20m
        Speaker: Antonio Nappi (CERN)
      • 17:10
        Oracle partner talk: Making databases smarter and faster: innovations enabled by engineering software and hardware together 20m

        We present examples that illustrate how jointly designing a database and its underlying hardware enables innovations that overcome substantial technological challenges. Some are fundamental advances in the state of the art, and all yield reliability and performance improvements that discrete component (“converged”) computer systems can rarely attain. For example, tailoring internal network protocols enables analytics queries to run without delaying OLTP commits; pushing computing into storage scales up throughput over limited bandwidth.

        This talk is for anyone interested in infrastructure-grade computer systems — not just databases — and aims to reveal the thinking of the technical staff at a large-scale development organization with a reputation for building and delivering systems that are critical for IT infrastructure across all sectors of the global economy.

        Speaker: Cris Pedregal (Oracle)
    • 17:35 19:35
      Cocktail and poster session Pas Perdus

      Pas Perdus

      CERN

      • 17:35
        Poster 1: Machine learning pipelines with Apache Spark and Intel Big DL 10m
        Speakers: Matteo Migliorini (Universita e INFN, Padova (IT)) , Viktor Khristenko (CERN)
      • 17:45
        Poster 2: DAQDB: a Key-value store for data acquisition 10m
        Speaker: Danilo Cicalese (CERN)
      • 17:55
        Poster 3: Monitoring JAVA application servers 10m

        Successful operations require constant awareness about the system state. Time to time, different obstacles show up, which have to be resolved as fast as possible, because user satisfaction highly depends on time. The prompt problem resolution depends on two steps: localization of the malfunctioning component and the recovery action. Monitoring aims to strike down the exploration time and provide a detailed information about the case. This poster provides an overview of Java applications monitoring used by the CERN IT-DB group.

        Speaker: Viktor Kozlovszky (CERN)
      • 18:05
        Poster 4: Physics data analysis and data reduction at scale with Apache Spark 10m
        Speaker: Vaggelis Motesnitsalis (CERN)
      • 18:15
        Poster 5: Physics data processing and machine learning in the Cloud 10m
        Speakers: Luca Canali (CERN) , Riccardo Castellotti (CERN)
  • Thursday, 24 January