SCIENCE BRINGS NATIONS TOGETHER
HybriLIT WORKSHOP 2026: From infrastructure to discovery

→ Europe/Moscow
MLIT-134/5-* - Конференц-зал (MLIT)

MLIT-134/5-* - Конференц-зал

MLIT

150
Дмитрий Подгайный (JINR)
Description

HYBRILIT WORKSHOP 2026 is the annual meeting of the HybriLIT user community, organized by the HybriLIT heterogeneous computing team.

Each year, the workshop brings together platform users, developers, and researchers to share experiences, discuss recent developments, and stay up to date with the evolving HybriLIT ecosystem and the Govorun supercomputer.

The workshop serves as a key opportunity to learn about new services, tools, and infrastructure updates, discover practical tips and efficient computational workflows, and explore emerging developments in high-performance computing, computational physics, and related research fields. Through talks and hands-on tutorials, participants gain practical knowledge, exchange ideas with the HybriLIT team and fellow users, and discover new opportunities to enhance their scientific computing workflows.

The workshop is organized by the HybriLIT team and is dedicated to the 70th anniversary of JINR and the 60th anniversary of MLIT.


Chairman:

Dmitry Podgainy, MLIT

Scientific Secretary:
Shushanik Torosyan, MLIT

Organizing Committee Members:
Oksana Streltsova, MLIT
Ilya Kalagin, MLIT
Maxim Zuev, MLIT
Maria Lyubimova, MLIT
Margarit Kirakosyan, MLIT
Stankus Daria, MLIT
Miroslav Iliaš, BLTP

Contacts:
Address:    MLIT Conference Hall
E-mail:       shushanik@jinr.ru
URL:
          https://indico.jinr.ru/e/hybrilit-workshop-2026

Important:
Participants are required to bring their personal laptops for the hands-on sessions.

Scientific secretary
Registration
Registration form
  • Thursday 26 November
    • 13:30 → 13:50
      HybriLIT Platform: a heterogeneous computing environment for scientific research 20m

      The presentation outlines the structure of the HybriLIT Platform, its software and information environment for computationally intensive tasks, the ML/DL/HPC ecosystem, and services for developing algorithms and software.

      Speakers: Дмитрий Подгайный (JINR), Оксана Стрельцова (JINR)
    • 13:50 → 14:10
      Getting started with HybriLIT: common pitfalls and best practices 20m
      Speakers: Shushanik Torosyan (LIT), Маргарит Киракосян (LIT)
    • 14:10 → 14:30
      Launching jobs through the SLURM resource manager 20m
      Speakers: Виктория Селянина (JINR), Дмитрий Беляков (MLIT JINR)
    • 14:30 → 14:40
      TBD 10m
      Speaker: Мария Любимова
    • 14:40 → 15:00
      Data storage systems of the HybriLIT educational and testing polygon and the “Govorun” supercomputer 20m

      The presentation will describe the data storage system of the HybriLIT training and testing facility and the “Govorun” supercomputer, which are part of the Multifunctional Information and Computing Complex (MICC) of the Laboratory of Information Technologies (LIT) at JINR. The “Govorun” supercomputer, which is part of the platform, features a hyperconverged architecture based on a hierarchical data model (hot, warm, and cold data tiers) and built on the Lustre and NFS/ZFS file systems to process data from the NICA experimental complex. Examples of specialized data storage configurations will be presented for users’ home directories, massively parallel computing, data processing for physics experiments, and radiobiological research.

      Speakers: Александр Кокорев, Дмитрий Беляков (MLIT JINR), Дмитрий Подгайный (JINR)
    • 15:00 → 15:20
      Lattice simulations of rotating and accelerated QCD on the HybriLIT Platform 20m
      Speaker: Dr Artem Roenko (BLTP, JINR)
    • 15:20 → 15:40
      Coffee Break 20m
    • 15:40 → 16:10
      HLIT-VDI: updating services for working with proprietary software 30m
      Speakers: Mikhail Matveyev (JINR), Shushanik Torosyan (LIT), Маргарит Киракосян (LIT)
    • 16:10 → 17:10
      TUTORIAL: TBD 1h

      The tutorial will demonstrate how to use the Mathematica software package in three modes: via the HLIT VDI service, within the ML/DL/HPC ecosystem, and in Batch mode. The following problems will be used as examples “Numerical Simulation of a Phase Transition in the Two-Dimensional Ising Model Using the Monte Carlo Method”, “The Problem of Controlling a Single-Stage Rocket During Its Injection into Near-Earth Orbit”

      Speakers: Anastasia Anikina (MLIT, JINR), Dmitrii Kokaev, Maxim Zuev (MLIT JINR), Mikhail Matveyev (JINR), Александр Хмелев (Студент), Маргарит Киракосян (LIT), Оксана Стрельцова (JINR), Юрий Калиновский (Joint Institute for Nuclear Research)
    • 17:10 → 18:10
      SciPy in Computational Physics: Experimental Data Fitting 1h

      This practical tutorial introduces data fitting and model evaluation using Python. Participants will learn to visualize measurements, estimate model parameters, and assess the quality of a fit through hands-on examples relevant to physics.

      Speakers: Adiba Rahmonova (Joint Institute for Nuclear Research), Сара Шадмехри
  • Friday 27 November
    • 14:00 → 14:20
      Quantum computing testbed for the HybriLIT heterogeneous platform 20m

      A quantum computing testbed has been deployed on the resources of the ML/DL/HPC ecosystem of the HybriLIT heterogeneous platform. The server’s specifications support computations on both CPUs and GPU accelerators. Example problems will be presented, and the impact of the computational resource configuration on performance will be demonstrated.

      Speakers: Maxim Zuev (MLIT JINR), Dr Yuri Palii (MLIT, JINR), Алла Боголюбская (JINR, Dubna), Дмитрий Беляков (MLIT JINR), Дмитрий Подгайный (JINR), Евгений Толочко, Оксана Стрельцова (JINR)
    • 14:20 → 14:50
      NDMSPC: Exploring Multi-Dimensional Space Analysis 30m

      The NDMSPC project introduces an innovative THnSparse analysis framework for high-dimensional data, directly addressing the challenge of fitting N-dimensional histograms within memory. Our solution organizes the THnSparse structure into sub-chunks, optimizing memory efficiency and enabling scalable processing. Beyond numerical values, the framework allows bin content to be arbitrary objects, greatly expanding data representation capabilities. A key feature is the ability to execute user-defined jobs for each bin, providing powerful and flexible custom analysis. This presentation will highlight the NDMSPC framework's design, benefits, and applications in advanced scientific data analysis.

      Speaker: Martin Vala (JINR)
    • 14:50 → 15:10
      Development of web services based on deep learning models on the HybriLIT platform 20m

      Using the development of services for automating the analysis of radiobiological data as an example, the presentation will demonstrate the capabilities of the HybriLIT heterogeneous platform form; annotating experimental data (images and videos) using the CVAT platform; training neural network models on the resources of the ML/DL/HPC ecosystem; developing and deploying web services based on trained models.

      Speakers: Maxim Zuev (MLIT JINR), Максим Бондарев (MLIT JINR), Оксана Стрельцова (JINR), Сара Шадмехри, Татевик Бежанян (LIT JINR)
    • 15:10 → 15:40
      Coffee Break 30m
    • 15:40 → 16:40
      TUTORIAL: TBD 1h
      Speaker: Оксана Стрельцова (JINR)
    • 16:40 → 17:00
      Comparative DFT Study of Adsorption of Superheavy Elements Cn, Fl and Mc, and Their Lighter Homologs on Au Surface 20m

      This study presents a Density Functional Theory (DFT) investigation of
      the adsorption of the superheavy elements (SHEs) Copernicium (Cn),
      Flerovium (Fl), and Moscovium (Mc), together with their lighter homologs
      Hg, Pb, and Bi, on Au(111) and Au(200) surfaces. The work aims to
      identify distinguishing features in the adsorption behavior of SHEs as a
      precursor to experimental studies. Calculations were performed using the
      PBE and PBEsol functionals with Optimized Norm-Conserving Vanderbilt
      Pseudopotentials (ONCVPSP), and the adsorption energetics and
      thermodynamics include rigorous spin-orbit coupling (SOC) and dispersion
      (DFT-D4) corrections. The results reveal systematic trends in binding
      strength and geometry, highlighting the relativistic and dispersion
      effects that shape the surface chemistry of superheavy elements, and
      provide a theoretical basis for future experimental characterization.

      Speakers: Dr Дипаян Сен (Bogoliubov Laboratory of Theoretical Physics, Joint Institute for Nuclear Research), Dr Мирослав Ильяш (BLTP JINR)
    • 17:00 → 17:10
      Benchmarking Machine-Learning Interatomic Potentials for the Adsorption of Atoms on a Gold Surface 10m
      Speaker: Neşet Koçkar
    • 17:10 → 17:25
      Computational Molecular Physics at Work 15m
      Speakers: Dr Дипаян Сен (Bogoliubov Laboratory of Theoretical Physics, Joint Institute for Nuclear Research), Dr Мирослав Ильяш (BLTP JINR)
    • 17:25 → 17:45
      Q&A session 20m