Focus on:  Hide Contributions
Back to Conference View
Login

NGT Algorithm Workshop – Lattice QCD at the large scale on exascale computing facilities

→ Europe/Zurich
4/3-006 - TH Conference Room (CERN)

4/3-006 - TH Conference Room

CERN

110
Show room on map
Description

Lattice Quantum Field Theory (LQFT) is a unique approach to control non perturbative physics. Driven by algorithmic and computational improvements, search for new physics with the support of LQFT methods is becoming possible. Namely with LQFT simulations including dynamical quarks at ever finer lattice spacings and larger physical box sizes per mille precision of standard model observables at the low energy regime can be reached.

The workshop is embedded within the Next-Generation-Triggers (NGT) grant, which pushes the intensity frontier of this endeavor by improving systematics, by utilizing novel algorithms, like machine learning techniques, and by optimizing software for novel computing architecture.

The kick-off workshop will highlight these challenges with the focus on lattice QCD calculations. Topics are techniques for variance reduction, like multi-level sampling, and gauge evolutions, like sampling algorithms to avoid topological freezing and the challenge of how to adapt to novel architectures, for example scientific computing in the era of large scale AI.

The aim of the workshop is to bring leading researchers in lattice QCD and computing experts from NGT working groups together. We plan 4-6 talks per day for three days. 

Speakers:
Ryan Abbott (MIT)
Simone Bacchio (Cyprus)
Lorenzo Barca (DESY)
Peter Boyle (BNL)
Claudio Bonanno (Madrid)
Timo Eichhorn (Wuppertal)
Jacob Finkenrath (CERN)
Roman Gruber (ETH)
Tim Harris (ETH)
Alessandro Lupo (Marseille)
Sofia Vallecorsa (CERN)
Michael Wagman (Fermilab)
Evan Weinberg (Nvidia)
Urs Wenger (Bern)
Yukari Yamauchi (University of Washington)

Organising committee:
Kate Clark
Matteo Di Carlo
Felix Erben
Jacob Finkenrath
Andreas Jüttner
Bartosz Kostrzewa
Simon Kuberski
Stefan Roiser
Phiala Shanahan

J. Tobias Tsang

Poster
Support
Participants
  • Alessandro Lupo
  • Alex Lasa Lamarca
  • Andreas Juttner
  • Claudio Bonanno
  • Enrique Rico Ortega
  • Felix Erben
  • Fernando Romero-López
  • Francesco Sanfilippo
  • Gianluca Fuwa
  • Isabel Campos Plasencia
  • Jacob Finkenrath
  • Jakob Hoffmann
  • Juan Fernández de la Garza
  • Kostas Orginos
  • Lorenzo Barca
  • Lukasz Michalski
  • Martin Lüscher
  • Michael Peardon
  • Michael Wagman
  • Orlando Oliveira
  • Patrick Fritzsch
  • Peter Boyle
  • Roman Gruber
  • Ryan Abbott
  • Simon Kuberski
  • Simone Bacchio
  • Stefan Roiser
  • Tim Harris
  • Timo Eichhorn
  • Tobias Tsang
  • Urs Wenger
  • Yukari Yamauchi
  • +6
NGT Algorithm Workshop
Zoom Meeting ID
61282747137
Host
Elena Gianolio
Alternative hosts
Tobias Tsang, Simon Kuberski, Matteo Di Carlo, Felix Benjamin Erben, Jacob Friedrich Finkenrath, Pascal Pignereau
Passcode
88154891
Useful links
Join via phone
Zoom URL
  • Monday 9 December
    • 10:00
      Coffee 4/2-011 - TH common room

      4/2-011 - TH common room

      CERN

      15
      Show room on map

      Prelude to the workshop

    • Variance reduction: Effective masses 4/3-006 - TH Conference Room

      4/3-006 - TH Conference Room

      CERN

      110
      Show room on map

      Long distance noise and excited states
      Q: Is multi-level sampling enough to increase long distance behavior ?

      • 1
        Introduction
        Speaker: Jacob Friedrich Finkenrath (CERN)
        Day_1.pdf
      • 2
        Spectral densities from the lattice
        Speaker: Alessandro Lupo (CPT, Aix-Marseille Université)
        main.pdf
      • 3
        Block Lanczos and signal to noise
        Speaker: Michael Wagman
        Wagman_CERN_24.pdf
    • Variance reduction: Multi Grid deflation and Multi Level sampling 4/3-006 - TH Conference Room

      4/3-006 - TH Conference Room

      CERN

      110
      Show room on map

      Long distance noise and excited states
      Q: Is multi-level sampling enough to increase long distance behavior ?

    • Machine learning at CERN 4/3-006 - TH Conference Room

      4/3-006 - TH Conference Room

      CERN

      110
      Show room on map
    • Reception 501 - CERN Restaurant 1

      501 - CERN Restaurant 1

      The workshop reception will take place in restaurant R1

  • Tuesday 10 December
    • Novel sampling algorithms for fine lattice spacings: Multi-tempering and Gauge flows 4/3-006 - TH Conference Room

      4/3-006 - TH Conference Room

      CERN

      110
      Show room on map

      Fine lattice spacings and ML for physics
      Q: Are local updates enough to move the topology at fine lattice spacings ?

      • 8
        Introduction Day 2 4/3-006 - TH Conference Room

        4/3-006 - TH Conference Room

        CERN

        110
        Show room on map
        Speaker: Jacob Friedrich Finkenrath (CERN)
        Day_2.pdf
      • 9
        Parallel tempered metadynamics 4/3-006 - TH Conference Room

        4/3-006 - TH Conference Room

        CERN

        110
        Show room on map
        Speaker: Timo Eichhorn
        NGT2024_Eichhorn.pdf
      • 10
        Full QCD simulations with milder topological freezing via Parallel Tempering on Boundary Conditions 4/3-006 - TH Conference Room

        4/3-006 - TH Conference Room

        CERN

        110
        Show room on map

        I will present a novel algorithm designed to mitigate the infamous topological freezing problem in full QCD lattice simulations, the Parallel Tempering on Boundary Conditions (PTBC), and discuss results obtained with PTBC in state-of-the-art simulations both at zero and finite temperature.

        Speaker: Claudio Bonanno (IFT UAM/CSIC Madrid)
        talk_CERN_Dec24_CBonanno.pdf
      • 10:35
        Coffee break 4/2-011 - TH common room (CERN)

        4/2-011 - TH common room

        CERN

        15
        Show room on map
      • 11
        Normalizing Flows for Gauge Theories 4/3-006 - TH Conference Room

        4/3-006 - TH Conference Room

        CERN

        110
        Show room on map
        Speaker: Ryan Abbott (MIT)
        AbbottRyanNGT2024.pdf
      • 12
        Continuous gauge flows 4/3-006 - TH Conference Room

        4/3-006 - TH Conference Room

        CERN

        110
        Show room on map
        Speaker: Simone Bacchio
        Bacchio - CNF.pdf
    • Novel sampling algorithms for fine lattice spacings: Novel Sampling and Discussions 4/3-006 - TH Conference Room

      4/3-006 - TH Conference Room

      CERN

      110
      Show room on map

      Fine lattice spacings and ML for physics
      Q: Are local updates enough to move the topology at fine lattice spacings ?

  • Wednesday 11 December
    • Novel computing architectures 4/3-006 - TH Conference Room

      4/3-006 - TH Conference Room

      CERN

      110
      Show room on map

      Scientific computing in the era of LLMs
      Q: How to utilize exascale machines optimal for LQFT ?

      • 16
        Computational and algorithmic advances using grid 4/3-006 - TH Conference Room

        4/3-006 - TH Conference Room

        CERN

        110
        Show room on map
        Speaker: Peter Boyle
        CERNtalk.pdf
      • 10:30
        Coffee break 4/2-011 - TH common room (CERN)

        4/2-011 - TH common room

        CERN

        15
        Show room on map
      • 17
        Computational and algorithmic advances using QUDA 4/3-006 - TH Conference Room

        4/3-006 - TH Conference Room

        CERN

        110
        Show room on map
        Speaker: Evan Weinberg (NVIDIA Corporation)
        weinberg_evan_ngt_cern_v4.pdf
      • 18
        Linkage of openQ*D and QUDA 4/3-006 - TH Conference Room

        4/3-006 - TH Conference Room

        CERN

        110
        Show room on map
        Speakers: Roman Gruber, Tim Harris (ETH Zürich)
        NGT_openqxd.pdf
    • Novel computing architectures: Novel computing architectures - Discussion and wrap up 4/3-006 - TH Conference Room

      4/3-006 - TH Conference Room

      CERN

      110
      Show room on map

      Scientific computing in the era of LLMs
      Q: How to utilize exascale machines optimal for LQFT ?

Powered by Indico v3.3.6-pre