Topics coveredQuality of software matters, whether you share it with others or not. Software should be easy to install, easy to use, and well documented. This training will cover those aspects from the perspective of the Python.ecosystem. However, it is also important that software is easy to maintain, so coding style matters, API-level documentation should be available, as well as a battery of tests to ensure the software's integrity. Of course, good design is at least as important.Subjects and Python modules that will be covered:user interface: argparse, configparserdocumentation: docstring, mkdocscode style and best practices, idiomatic Python: flake8, pylintunit testing: pytestobject oriented programming in Pythondesign patterns

This course is part of the "LRZ AI Training Series", a series of courses aiming at the needs and expectations of data analytics, big data & AI users at LRZ.Contents:Businesses worldwide are using artificial intelligence to solve their greatest challenges. Healthcare professionals use AI to enable more accurate, faster diagnoses in patients. Retail businesses use it to offer personalised customer shopping experiences. Automakers use it to make personal vehicles, shared mobility, and delivery services safer and more efficient. Deep learning is a powerful AI approach that uses multi-layered artificial neural networks to deliver state-of-the-art accuracy in tasks such as object detection, speech recognition, and language translation. Using deep learning, computers can learn and recognise patterns from data that are considered too complex or subtle for expert-written software.In this course, you’ll learn how deep learning works through hands-on exercises in computer vision and natural language processing. You’ll train deep learning models from scratch, learning tools and tricks to achieve highly accurate results. You’ll also learn to leverage freely available, state-of-the-art pre-trained models to save time and get your deep learning application up and running quickly.The course is co-organised by LRZ and NVIDIA Deep Learning Institute (DLI). All instructors are NVIDIA certified University Ambassadors.Learning Objectives:By participating in this course, you will:Learn the fundamental techniques and tools required to train a deep learning modelGain experience with common deep learning data types and model architecturesEnhance datasets through data augmentation to improve model accuracyLeverage transfer learning between models to achieve efficient results with less data and computationBuild confidence to take on your own project with a modern deep learning framework

Interactive exploration and analysis of large amounts of data from scientific simulations, in-situ visualization and application control are convincing scenarios for explorative sciences. Based on the open source software Jupyter or JupyterLab, a way has been available for some time now that combines interactive with reproducible computing while at the same time meeting the challenges of support for the wide range of different software workflows.Even on supercomputers, the method enables the creation of documents that combine live code with narrative text, mathematical equations, visualizations, interactive controls, and other extensive output. However, a number of challenges must be mastered in order to make existing workflows ready for interactive high-performance computing. With so many possibilities, it's easy to lose sight of the big picture. This course provides a detailed introduction to interactive high-performance computing.The following topics are covered:Introduction to JupyterParallel computing using JupyterInteractive & in-situ visualizationFrom ipywidgets to dashboards

AnnotationThis training introduces participants to Python for high-performance computing, covering parallel programming, performance optimization, and HPC resource utilization. Designed for researchers and developers, the course includes hands-on sessions to enhance practical skills.Target Audience and Purpose of the Course:Python's role in HPC and performance optimizationParallel programming techniques for efficient computingHow to utilize HPC resources effectivelyHands-on experience with lab exercises for practical skills Participants will have access to the Karolina supercomputer for hands-on sessions, utilizing both CPU and GPU resources. Karolina, operational since 2021, is the most powerful supercomputer in the Czech Republic and ranks among Europe's top systems. It features a standard part with 720 nodes, delivering 11.6 PFlop/s for traditional HPC simulations, and an accelerated section comprising 72 servers, each equipped with 8 GPU accelerators, achieving up to 360 PFlop/s for AI computations. This infrastructure supports complex scientific and industrial challenges, including numerical simulations, data analysis, and artificial intelligence applications.Level70% beginner, 30% intermediateLanguageEnglishPrerequisitesbeginner experience with programming in PythonTechnical requirements: Python and it’s dependenciesJupyter Notebook for interactive codingAnaconda (optional) for managing dependenciesTutorsTomas Martinovic is a senior researcher at the Advanced Data Analysis and Simulation Laboratory within the IT4Innovations National Supercomputing Center. His work primarily focuses on the data science, data visualisation, and mathematical modeling leveraging statistical methods and deep neural networks.Ghaith Chaabane Researcher at the Advanced Data Analysis and Simulation Laboratory within the IT4Innovations National Supercomputing Center.

Schedule 29/04/2025: 09.00 - 13.00 30/04/2025: 09.00 - 13.00 Topics coveredOpenMP is the de facto standard for shared memory programming in a scientific context. Using annotations, the C/C++/Fortran programmer can parallelize his code step by step. However, to obtain efficient code, it is mandatory to have a thorough understanding of the features and pitfalls of OpenMP. This course deals with those in detail. This course covers the OpenMP 3.1 standard, and gives an overview of the new features in OpenMP 4.0.Overview and execution modelWork sharing directivesData environmentPitfallsVerification of OpenMP applicationsOpenMP 4.0 overviewTarget audienceThis info session is primarily targeted at VSC-users, although other interested parties are welcome as well.Previous knowledgeGood working knowledge of C/C++ or Fortran programming is expected.Basic knowledge of Linux usage and HPC systems helps.Level: IntermediateResult/ObjectivesParticipants are able to start developing multicore applications using OpenMP.Good understanding of the semantics of OpenMP featuresAwareness of potential pitfallsFamiliarity with the new features in the OpenMP 4 standardRemarkThis course is based on a PATC course developed by dr. Rolf Rabenseifner (HLRS, Stuttgart).

This course will give a deep dive into the AMD InstinctTM GPU architecture and its ROCmTM ecosystem, including the tools to develop or port HPC or AI applications to AMD GPUs. Participants will be introduced to the programming models for the MI200 series GPUs and MI300A APU.The new unified memory programming model makes writing HPC applications much easier for a wide range of GPU programming models. We will cover how to use pragma-based languages such as OpenMP, the basic GPU programming language HIP, and performance portable languages such as Kokkos and RAJA. In addition, there will be presentations on other important topics such as GPU-aware MPI, and Affinity. The AMD tool suite, including the debugger, rocgdb, and the profiling tools rocprof, rocprof-sys, and rocprof-compute will also be covered.A short introduction will be given into the AMD Machine Learning software stack including PyTorch, JAX and Tensorflow and how they have been used in HPC.

What if you no longer have to install a broad range of scientific software from scratch on every laptop, HPC cluster, or cloud instance you use or maintain, without compromising on performance?The European Environment for Scientific Software Installations (EESSI, https://eessi.io) comes to the rescue!In this webinar series we will provide a comprehensive overview of EESSI: why we started it, how it works, how you can use it, ...You can register for the sessions listed below (either all of them, or selected ones).All sessions will be recorded. Recordings, slides, and materials used will be made publicly available shortly after each session via this page.In you have any questions regarding these webinars, please send an email to support@eessi.io.SessionsMonday 5 May 2025 (13:30-15:30 CEST): Introduction to EESSIMonday 12 May 2025 (13:30-15:30 CEST): Introduction to CernVM-FSMonday 19 May 2025 (13:30-15:30 CEST): Introduction to EasyBuild, incl. highlighted features in EasyBuild 5.0.0Monday 26 May 2025 (13:30-15:30 CEST): EESSI for CI/CDMonday 2 June 2025 (13:30-15:30 CEST): Using EESSI as the base for a system stackFormatOnline webinars (via Zoom)Mix of presentation & hands-on demos: ~1.5h of content, ~30min for Q&ARegistrationAttendance is free of cost, but registration is required.Register via https://event.ugent.be/registration/eessi202505Useful linksEESSI website: https://eessi.ioEESSI documentation: https://eessi.io/docsCernVM-FS website: https://cernvm.cern.ch/fsCernVM-FS documentation: https://cvmfs.readthedocs.ioEasyBuild website: https://easybuild.ioEasyBuild documentation: https://docs.easybuild.ioMultiXscale website: https://www.multixscale.eu

Schedule 06/05/2025: 9 a.m.-17.00 p.m.07​​​/05/2025: 9 a.m.-17.00 p.m. ContentThe Message Passing Interface is the de facto standard for distributed programming in a scientific context. It is implemented in libraries that can be used from C/C++/Fortran, although wrappers for other languages exist as well.To write efficient code, a thorough understanding of the semantics of the API is required, this course will emphasize potential performance hazards and pitfalls. The strengths and weaknesses of various MPI features are discussed. In several hands-on sessions, the participant has the opportunity to experiment.Parallel architectures and programming models MPI overviewMPI programming modelMessages and point-to-point communicationNon-blocking communicationDerived data typesVirtual topologiesCollective communicationMPI-2 overviewMPI-2 one-sided communicationParallel debuggingPI-2 parallel fileMPI-3 overviewThis MPI-course covers the full MPI-3 standard Learning objectives Participants are able to start developing distributed applications using MPI Good understanding of the semantics of MPI features Awareness of potential pitfalls Familiarity with the new features in the MPI-3 standard Remark This course material was developed by dr. Rolf Rabenseifner (HLRS, Stuttgart, Germany).

Over the past decade or more, most of the software that uses GPGPUs has been developed in CUDA since only NVIDIA provided such hardware.However, currently, at least three vendors provide GPU accelerators, AMD, Intel and NVIDIA. This implies that you need a vendor-agnostic programming model to use GPUs. OpenMPis such a model. It has long been used for shared-memory parallelism, but it has been extended to support offloading to accelerators.This training will show you how to use OpenMP to offload computations to GPUs. Learning objectives understand the principles of offloading computations to GPUs using OpenMP; know how to transfer data between the host and the device; be able to write programs that offload computations to GPUs. This training if for you if you want to develop software that uses GPUs, and want the software to be portable across GPUs of different vendors.

Topics covered All code contains bugs, finding and fixing them is boring. In this training, best practices are presented to reduce the number of bugs in your code. You will also learn about debuggers and debugging techniques to find bugs more efficiently. best practices in coding, code style, error handling using compiler options to warn about potential issues unit and functional testing using a debugger (breakpoints, inspection, watchpoints, tracing ...) zoo of bugs optional: debugging parallel code Previous knowledge experience programming C, C++, or Fortran familiarity with the bash command line optional: experience using MPI and/or OpenMP

This course will take place as an online event and is exclusively targeted for users of the Jülich supercomputers. The link to the online platform will be provided to the accepted registrants only.Contents:Disclaimer: This course offers the opportunity to attend lectures selectively, based on individual needs and knowledge levels. Participants are not required to attend all sessions as some may cover advanced or basic material. While flexibility is offered, it is important to assess personal needs and choose sessions accordingly, as attending lectures out of order may result in knowledge gaps.Research Centre Jülich provides cutting-edge high-performance computing resources to scientific groups and industry partners across Germany and Europe via the John von Neumann Institute for Computing. To help new users of JSC's supercomputers efficiently leverage their allocated resources, we offer an introductory course that covers system basics and best practices.The course includes theoretical lectures held every afternoon from Tuesday to Friday, and hands-on offered in the mornings from Wednesday to Friday. The hands-on are based on the previous afternoon's lectures and allow participants to put theory into practice. The hands-on sessions cover a wide range of topics, from basic login procedures to advanced techniques. For the more experienced, there's the option of bringing your own code or discussing current challenges with JSC experts, ensuring a tailored and interactive learning experience.Topics covered include:User account management with the JuDoor portalSystem access via SSH, Jupyter, and UNICORESystem configuration for JURECA and JUWELSFile systems, I/O, and data managementSoftware modules (compilers, MPI, math libraries, applications, debuggers, tools)Building software from sourceSubmitting jobs via the resource managerUsing GPUsPerformance tuningDeep LearningVisualization

This advanced C++ course provides both advanced C++ programming techniques and software design insight to help developers to create professional, high-quality code. The course provides deeper insight into C++ templates (type traits, SFINAE, C++20 concepts and forwarding references) and advice on how to create safer and cleaner function and class interfaces. Additionally, it provides an introduction to software design. For that purpose, it demonstrates the gravity of bad dependency management and shows how to properly reduce dependencies by means of design patterns.

C++ is widely used in many contexts. This training focusses on software engineering aspects of C++ such as building software, linting, testing libraries and applications, and so on. Also, best practices are discussed such as design patterns. When you complete this training you will be able to: assess code quality; build software using CMake; perform static code analysis and interprete the results; write unit tests using Catch2; write functional tests using CTest; use C++ package managagers such as Conan or vcpkg; use some design patterns that are useful in scientific programming; program in an efficient functional style using STL algorithms and rangs. Previous knowledge You will need experience programming in C++, this is not a training that teaches you how to program in C++. You also need to be familiar with the Bash command line.

Pandas solves the full stack of data analysis in Python; reading-in of data, mangling and manipulation, analysis, and visualisation (and much more, actually). It builds up on established Python packages and can be used interchangeably with them (like Numpy, matplotlib); it fits perfectly into the Jupyter Notebooks workflow of modern-day data analysis.This course will introduce Pandas with simple examples in hands-on exercises, highlighting the capabilities of the software package on the way with increasing complexity.Attendees should have run a Python program before and know the bare minimum of the programming language.You will be using your own computer for the exercises.

The workshop on High-Performance Computing (HPC) Visualization will focusing on cutting-edge tools and techniques for handling the vast datasets produced by large-scale simulations and observational surveys in astrophysics and cosmology.The growing complexity of astrophysical and cosmological data—whether from massive simulations or large-scale observational campaigns—requires powerful visualization tools to extract scientific insights. However, implementing these tools efficiently on diverse HPC platforms presents challenges such as hardware and software compatibility, scalability, and resource optimization.This workshop will explore strategies to overcome these challenges and highlight recent advances in high-performance visualization, including:Distributed visualization in heterogeneous architecturesParallel cinematic volume visualizationIn-situ visualizationVisualization methods for machine learning-assisted analysisWhile the primary focus is on astrophysics and cosmology, these techniques have broad applications across other scientific domains dealing with large and complex datasets. We encourage participation from researchers and developers interested in state-of-the-art visualization methods.This workshop is an initiative of the SPACE Centre of Excellence, an EU-funded project dedicated to enabling flagship European Astrophysics and Cosmology (A&C) simulation codes to run efficiently on pre-exascale and exascale architectures. Within this framework, we aim to develop advanced software solutions for data access, storage, and analysis, ensuring that HPC visualization keeps pace with the increasing complexity of simulation-driven research.The High-Performance Computing Visualization Workshop, organized with the support of NCC Czechia and CASTIEL2, will be a hybrid event, held online via Zoom and on-site at the Barcelona Supercomputing Center. Before the event, technical details will be sent to the accepted registrants.The on-site capacity is limited to 25 participants, online participation is unlimited. The course is open to participants from the Member States (MS) of the European Union (EU) and Associated/Other Countries to the Horizon Europe programme.Participation in the course is free of charge; however, participants must arrange travel and accommodation at their own expense. Registered participants will receive a list of recommended hotels, including special room rates.

As the diversity of hardware accelerators increases, and CPUs keep adding more cores, it becomes increasingly important to write software that can run in a heterogeneous environment. This training will introduce you to Kokkos, a programming model that allows you to write code that can run on CPUs, GPUs, and other accelerators. Kokkos is a C++ library that provides abstractions for parallel programming, and can generate code for different hardware platforms. Learning Objectives When you complete this training, you will: understand the principles of the Kokkos programming model; be able to write parallel code that can run on CPUs, GPUs, and other accelerators; know about relevant execution patterns and how to implement them with Kokkos; be able to handle data transfers between the host and the device. This training is for you if you want to develop software that can run on CPUs, GPUs, and and other accelerators.

This course is part of the "LRZ AI Training Series", a series of courses aiming at the needs and expectations of data analytics, big data & AI users at LRZ.The course is organised as an on-site even at LRZ in Garching near Munich. There will be no possibility to join online remotely via video conference. Participants are expected to bring their own laptops running the latest version of Chrome or Firefox. There are no PCs installed in the course room!Contents:Modern deep learning challenges leverage increasingly larger datasets and more complex models. As a result, significant computational power is required to train models effectively and efficiently. Learning to distribute data across multiple GPUs during deep learning model training makes possible an incredible wealth of new applications utilizing deep learning.Additionally, the effective use of systems with multiple GPUs reduces training time, allowing for faster application development and much faster iteration cycles. Teams who are able to perform training using multiple GPUs will have an edge, building models trained on more data in shorter periods of time and with greater engineer productivity.This workshop teaches you techniques for data-parallel deep learning training on multiple GPUs to shorten the training time required for data-intensive applications. Working with deep learning tools, frameworks, and workflows to perform neural network training, you’ll learn how to decrease model training time by distributing data to multiple GPUs, while retaining the accuracy of training on a single GPU.The course is co-organised by LRZ and NVIDIA Deep Learning Institute (DLI). All instructors are NVIDIA certified University Ambassadors.Learning Objectives:By participating in this workshop, you’ll:Understand how data parallel deep learning training is performed using multiple GPUsAchieve maximum throughput when training, for the best use of multiple GPUsDistribute training to multiple GPUs using Pytorch Distributed Data ParallelUnderstand and utilize algorithmic considerations specific to multi-GPU training performance and accuracy

Do you want to automate your SURF Research Cloud workflows using the API? Join this training to learn how to use the API to prepare environments for a group of users, control workspaces, and connect services. This is a follow-up to the Introduction to SURF Research Cloud training.Sign upFor whom?This course is designed for researchers and IT professionals who are interested in automating their workflows and exploring the capabilities of SURF Research Cloud through API-driven solutions. The programming language used in the training is Python. The language of instruction is English.RequirementsBring your own laptop.A browser (Chrome or Firefox is good enough).Experience in using SURF Research Cloud (preferably knowledge from Introduction to SURF Research Cloud training).Prior knowledge of basic programming, command line tools, cloud computing concepts, and an API would be beneficial.Sign up now Advanced Use of SURF Research Cloud (API)

C++ is a multi-paradigm programming language supporting procedural, object-oriented, generic and functional programming styles. This course will provide a practical introduction to C++, adhering to the latest official language standard at the time of the course.The participants will study small example programs, each demonstrating a certain aspect of C++, and then do simple programming exercises using the lessons learned from the examples. The initial focus of the course will be to make the participants comfortable utilizing modern C++, e. g., solving small problems using the STL containers and algorithms along with lambda functions. Syntax will be explained in detail when needed. Once the participants are familiar and comfortable with the easy-to-use aspects of modern C++, the powerful abstraction mechanisms of the language, such as classes and class hierarchies, and templates will be presented at depth. It is hoped that this course will encourage fruitful application of the programming language and provide a good foundation for further learning.It is assumed that the participants have previous programming experience in languages such as C, C++, Python, Java and Fortran. This course introduces programming in C++20. It is not meant to be a beginners' introduction to programming.

May 19-23 2025 @ University of Modena and Reggio EmiliaVia Campi 213/BModena, ItalyApplication deadline March 31[00] OverviewThis school is designed for researchers interested in advanced computational methods for studying light-matter interactions, crucial for optoelectronic devices, quantum computing, and energy applications.Participants will receive introductory and advanced lectures on many-body perturbation theory (MBPT), covering topics such as the GW approximation, Bethe-Salpeter equation (BSE), non-linear optics, and recent algorithmic advances for 2D systems and metals.Hands-on sessions will focus on practical simulations using the YAMBO code within a high-performance computing (HPC) environment, including GPU-accelerated machines. Additionally, participants will learn Python-based post-processing and data analysis with YamboPy.YAMBO is a flagship code of the MaX Centre of Excellence, ICSC PNRR Italian National Centre for HPC, Big Data and Quantum Computing and the Hanami project.

RationaleThe Quantum opEn-Source Package for Research in Electronic Structure, Simulation, and Optimization, best known as Quantum Espresso (QE), is a software suit for quantum-mechanical materials modeling and is distributed as a free software under the GNU General Public License. QE is an open initiative in collaboration with many known international groups (SISSA, ICTP, CINECA, EPFL, Oden Institute and University of Texas at Austin) and is coordinated by the Quantum ESPRESSO Foundation. The suit is based on the framework of density functional theory, by resorting to plane-wave basis-sets and different datasets for pseudopotentials (norm-conserving, ultrasoft and PAW). QE can be employed to compute electronic ground-state calculations, structural optimizations, ab-initio molecular dynamics, response properties (DFPT), spectroscopic properties, quantum transport, pseudopotential generation.The main core of the QE software is designed to efficiently exploit the architecture of current HPC systems and developed for the exascale computing. The parallelization is achieved by resorting to multi-core parallelization (both MPI and OpenMP) and GPU acceleration, allowing distribution of calculations and data structure across processors and/or GPU porting. This workshop will provide the opportunity for researchers in computational chemistry and materials science to share their work, and discuss, about using the QE suit, not only in the applications that can be done by using this suite of codes, but also best practices for efficient exploitation over HPC systems, and limitations of the code. Scientific TopicsThe meeting program will cover several fundamental topics in the field employing different theoretical methodologies which are best fitted for the properties to be probed and implemented in the Quantum Espresso software package:Phonon frequencySpectroscopic propertiesElectron-phonon couplingHubbard U and magnetismBerry phase polarizationMolecular dynamics and potencial energy surfacesElectrochemistryGIPAW: NMR and hyperfine interactions

ContentFortran is used a lot in the context of HPC. For many, it has a reputation of being an old and ugly programming language. However, modern Fortran is a far cry from Fortran 77 that its detractors have in mind. Fortran 2003/2008 is a language that is well tailored towards scientific computing. This training introduces the language and its features. Learning Objectives When you complete this training, you will be able to read and understand well-written Fortran code; write intermediate-level Fortran programs; have a good starting point to familiarize yourself with the more sophisticated aspects of Fortran by practicing; know about best practices and pitfalls for Fortran programmers.

The End-to-End AI for Science Bootcamp will be hosted virtually for two half days on May 27-28. This bootcamp provides a step-by-step overview of the fundamentals of deep neural networks, walks attendees through the hands-on experience of building and improving deep learning models using a framework that uses the fundamental laws of physics to model the behaviour of complex systems, and enables attendees to visualize the outputs of the trained model.This online bootcamp is a hands-on learning experience where you will be guided through step-by-step instructions with teaching assistants on hand to help throughout.This bootcamp is co-organized by the Vienna Scientific Cluster (VSC), IT4Innovations National Supercomputing Center (IT4I), High-Performance Computing Center Stuttgart (HLRS), Jülich Supercomputing Centre (JSC), Leibniz Supercomputing Centre (LRZ), University of Donja Gorica (UDG), Academic Computer Centre Cyfronet AGH (Cyfronet), Linköping University (LiU), Research Institutes of Sweden (RISE), HPC Vega at IZUM (IZUM), OpenACC organization, and NVIDIA for EuroCC Austria, EuroCC Czechia, EuroCC@GCS, EuroCC Montenegro, EuroCC Poland, EuroCC Sweden, and EuroCC Slovenia, all National Competence Centres for High-Performance Computing.

ContentLarge Language Models (LLMs) are a class of machine learning models that have recently gained a lot of attention. These models are trained on large amounts of data and after training can be used in many applications. Although models from OpenAI and Google can be used as services online, it is often desirable to have a model that can be used offline. This training will show you how to deploy and use such models locally.Learning Objectivesunderstand what LLMs are and how they are trained; be able to use a pre-trained LLM for text generation; be able to use Retrieval Augmented Generation (RAG) for question answering on your own data;understand how quantization works and how it can be used to reduce the size of a model; be able to fine-tune a pre-trained LLM for a specific task. Target audience This training is for you if you need to deploy Large Language Models (LLMs) on your own infrastructure. Prior knowledgeYou will need experience programming in Python. This is not a training that starts from scratch.Familiarity with Linux or HPC environments is recommended.

Advanced SIESTA Workshop 2025June 2, 2025 - June 5, 2025Registration deadline: May 5, 2025Location: Catalan Institute of Nanoscience and Nanotechnology, UAB Campus, Barcelona Area, SpainHosting node: CECAM-ESAdditional sponsors: Psi-k and MaX CoE OrganisersCatalina Coll (Institut Català de Nanociència i Nanotecnologia)José María Escartín Esteban (Catalan Institute of Nanoscience and Nanotechnology (ICN2))Roberta Farris (Institut Català de Nanociència i Nanotecnologia (ICN2))Federico Nicolás Pedron (Catalan Institute of Nanoscience and Nanotechnology (ICN2))Miguel Pruneda (Consejo Superior de Investigaciones Científicas (CSIC))The unique advantages of the SIESTA method[1][2] allow for the exploration of boundaries that are hard to be achieved by other DFT approaches. In this Workshop, we will showcase and discuss state-of-the-art research that relies on SIESTA and its interfaces with other computational tools.Examples of these include the calculation of advanced magnetic properties (relying on TB2J[3]), thermal conductivity (via TDEP[4]), electrochemical processes (via SIESTA's own QM/MM interface[5]), potential energy surface explorations (with SIESTA/LUA) and electron-phonon coupling phenomena (with the elph interface). The usage of SIESTA in exascale and pre-exascale systems will be discussed, covering different technologies available for the calculation of electronic properties with tens of thousands of atoms. We will also include a discussion on how machine learning technologies intertwine with SIESTA, be it as a means to accelerate and improve calculations, or to use SIESTA as a provider for training data in other neural network approaches.In this 3-day workshop, experts on the different topics will showcase current advances on their respective areas while also providing guidelines and opening the discussion for new approaches and ideas. A few selected sessions will also include short hands-on sections to further exemplify their capabilities.References[1] J. Soler, E. Artacho, J. Gale, A. García, J. Junquera, P. Ordejón, D. Sánchez-Portal, J. Phys.: Condens. Matter, 14, 2745-2779 (2002)[2] A. García, N. Papior, A. Akhtar, E. Artacho, V. Blum, E. Bosoni, P. Brandimarte, M. Brandbyge, J. Cerdá, F. Corsetti, R. Cuadrado, V. Dikan, J. Ferrer, J. Gale, P. García-Fernández, V. García-Suárez, S. García, G. Huhs, S. Illera, R. Korytár, P. Koval, I. Lebedeva, L. Lin, P. López-Tarifa, S. Mayo, S. Mohr, P. Ordejón, A. Postnikov, Y. Pouillon, M. Pruneda, R. Robles, D. Sánchez-Portal, J. Soler, R. Ullah, V. Yu, J. Junquera, The Journal of Chemical Physics, 152, (2020)[3] X. He, N. Helbig, M. Verstraete, E. Bousquet, Computer Physics Communications, 264, 107938 (2021)[4] F. Knoop, N. Shulumba, A. Castellano, J. Batista, R. Farris, M. Verstraete, M. Heine, D. Broido, D. Kim, J. Klarbring, I. Abrikosov, S. Simak, O. Hellman, JOSS., 9, 6150 (2024)[5] C. Sanz-Navarro, R. Grima, A. García, E. Bea, A. Soba, J. Cela, P. Ordejón, Theor. Chem. Acc., 128, 825-833 (2010)

Businesses today generate vast amounts of voice and text data from various channels, such as emails, news articles, calls, and customer reviews. This data holds valuable insights that can help companies uncover hidden patterns and key features. To harness this potential, the data needs to be efficiently processed and modeled using Natural Language Processing (NLP) algorithms.The goal of this course is to provide a fundamental understanding of NLP technology, which has attracted significant attention in fields such as marketing, customer service, and e-commerce. The course covers a range of algorithms and topics, including sentiment analysis, topic modeling, semantic search, chatbots, transformers, and LLM fine-tuning. The algorithms are illustrated with figures, diagrams, and practical examples to show how you can apply these techniques to enhance your business operations.

The summer school will include lectures and hands-on sessions on the following topics (subject to change):Molecular dynamics simulationsBiomolecular dockingFree energy calculationsAdvanced sampling methodsBioExcel Building Blocks (BioBB)The summer school will include lectures on the following topics (subject to change):Quantum mechanics/molecular mechanics (QM/MM)Machine learningDuring the hands-on practicals you will learn how to use the BioExcel flagship software: GROMACS, HADDOCK, PMX and BioBB. The trainers, developers and experts will provide guidance and support in using the software. The School is also a great opportunity to discuss your own research and build your professional network with other participants.Target audience and prerequisitesThe BioExcel School is intended primarily for PhD and post-doctoral researchers using or planning to use biomolecular modeling and simulation in their everyday research.Familiarity with Linux and basic knowledge of molecular modelling software is a requirement.We can only accept participants from EuroHPC Joint Undertaking member institutions (unfortunately excluding the UK).We highly encourage applications from industrial entities, women and underrepresented groups.

Registration deadline: March 28, 2025Location: Zadar, CroatiaHosting node: CECAM-HQAdditional sponsors: Psi-k, MaX CoE, Cost Action DAEMON, University of Zadar, Ruđer Bošković Institute, Croatian Science FoundationOrganisersKatarina Batalović (VINCA Institute, University of Belgrade)Federico Grasselli (University of Modena and Reggio Emilia)Ivor Loncaric (Rudjer Boskovic Institute)Juraj Ovčar (SISSA)Kevin Rossi (TU Delft)In the last few years, machine learning and data-driven approaches, in synergy with established atomistic and molecular modeling methods, are enabling a paradigm shift in the accuracy and informativeness of computational predictions. [1, 2, 3] In this regard, we are moving from a moment of emergence to one of consolidation and maturity. This statement is corroborated not only in the remarkable growth of investigation exploiting machine learning methods, but also by the uptake of these methods in industries across the pharma, specialty chemicals, energy, and aerospace, as well as in the strong support from the market of start-ups promising the discovery and synthesis of materials and molecules with advanced properties thanks to artificial intelligence and physics-based modeling.In this regard popularizing robust high-throughput workflows, established data analytics approaches, and machine learning-enabled realistic atomistic simulations pipelines to new generation of researchers approaching the field is expected to act as a level playing field and to accelerate discoveries and innovation in the domain of computational materials science, chemical physics, physical chemistry, and computational biophysics.To this end, our events aims at providing a pedagogical introduction to young researchers, as well as established ones interested in rapidly adopting machine learning methods in their work. In particular, we have identified 8 topics that are nowadays commonly part of investigation concerning the understanding, characterization, or design of molecules, materials, and processes, namely:1. Machine Learning accelerated High-throughput searches based on density functional theory (and beyond) simulations [4]2. Bayesian optimization of materials and molecular properties [5]3. Generative models for materials and molecules design [6]4. Large language models and embeddings for materials and molecules property prediction [7]5. Automated construction of machine learning interatomic potentials [8]6. Non-adiabatic and excited state dynamics with machine learning models [9]7. Learning coarse-grained models [10]8. Integrating experimental data in machine-learning-driven materials discovery [11]The first block concerns the exploration of large chemical spaces, generally accounting for equilibrium or ensemble averaged information, utilizing an array of methods, which ranges from materials and chemo-informatics approaches, to other closer to artificial intelligence. The second block of lecture reflects the key role of dynamical properties in determining the properties of a material, a molecule, or an experimental measurement. Also for this case, a diverse set of methodologies is considered, also in relation to a multi-scale vision.The school will cover each of these topics through an introductory lecture and hands-on tutorial showcasing a realistic scenario. Importantly, the location of the school (Zadar, Croatia), is expected to enable the transfer of knowledge and cross-fertilization of network of researchers from the so-called (in the EU nomenclature) widening or inclusiveness countries in Europe.Participants are strongly encouraged to submit an oral or poster contribution to the event by specifying this in the CECAM application form, accessible after clicking "Participate" on this webpage. In the application form, please include the contribution title and abstract in the "Your message" section. Several slots for oral contributions are available, and applicants will be informed about their acceptance for attendance and the type of contribution by April 17, 2025.The school is free of charge. Travel and accommodation reimbursements are available for members of the COST DAEMON Action, in accordance with the official rules outlined in the Annotated Rules for COST Actions. Applicants intending to apply for this funding should clearly indicate this in their application form. To facilitate the organization of coffee breaks, refreshments, and the social dinner, we ask participants to also indicate any dietary restrictions in the application form.References[1] S. Bauer, P. Benner, T. Bereau, V. Blum, M. Boley, C. Carbogno, C. Catlow, G. Dehm, S. Eibl, R. Ernstorfer, Á. Fekete, L. Foppa, P. Fratzl, C. Freysoldt, B. Gault, L. Ghiringhelli, S. Giri, A. Gladyshev, P. Goyal, J. Hattrick-Simpers, L. Kabalan, P. Karpov, M. Khorrami, C. Koch, S. Kokott, T. Kosch, I. Kowalec, K. Kremer, A. Leitherer, Y. Li, C. Liebscher, A. Logsdail, Z. Lu, F. Luong, A. Marek, F. Merz, J. Mianroodi, J. Neugebauer, Z. Pei, T. Purcell, D. Raabe, M. Rampp, M. Rossi, J. Rost, J. Saal, U. Saalmann, K. Sasidhar, A. Saxena, L. Sbailò, M. Scheidgen, M. Schloz, D. Schmidt, S. Teshuva, A. Trunschke, Y. Wei, G. Weikum, R. Xian, Y. Yao, J. Yin, M. Zhao, M. Scheffler, Modelling Simul. Mater. Sci. Eng., 32, 063301 (2024)[2] S. Back, A. Aspuru-Guzik, M. Ceriotti, G. Gryn'ova, B. Grzybowski, G. Gu, J. Hein, K. Hippalgaonkar, R. Hormázabal, Y. Jung, S. Kim, W. Kim, S. Moosavi, J. Noh, C. Park, J. Schrier, P. Schwaller, K. Tsuda, T. Vegge, O. von Lilienfeld, A. Walsh, Digital Discovery, 3, 23-33 (2024)[3] J. Margraf, Angew. Chem. Int. Ed., 62, (2023)[4] J. Schmidt, J. Shi, P. Borlido, L. Chen, S. Botti, M. Marques, Chem. Mater., 29, 5090-5103 (2017)[5] M. Todorović, M. Gutmann, J. Corander, P. Rinke, npj. Comput. Mater., 5, 35 (2019)[6] R. Gómez-Bombarelli, J. Aguilera-Iparraguirre, T. Hirzel, D. Duvenaud, D. Maclaurin, M. Blood-Forsythe, H. Chae, M. Einzinger, D. Ha, T. Wu, G. Markopoulos, S. Jeon, H. Kang, H. Miyazaki, M. Numata, S. Kim, W. Huang, S. Hong, M. Baldo, R. Adams, A. Aspuru-Guzik, Nature. Mater., 15, 1120-1127 (2016)[7] A. M. Bran, S. Cox, O. Schilter, C. Baldassari, A. White, P. Schwaller, Nat. Mach. Intell., 6, 525-535 (2024)[8] Y. Yang, S. Zhang, K. Ranasinghe, O. Isayev, A. Roitberg, Annual Review of Physical Chemistry, 75, 371-395 (2024)[9] J. Westermayr, P. Marquetand, Chem. Rev., 121, 9873-9926 (2020)[10] B. Mohr, D. van der Mast, T. Bereau, J. Chem. Theory Comput., 19, 4770-4779 (2023)[11] R. Miyazaki, K. Belthle, H. Tüysüz, L. Foppa, M. Scheffler, J. Am. Chem. Soc., 146, 5433-5444 (2024)

HPC application developers encounter significant challenges getting their codes to run correctly on leadership computer systems consisting of large numbers of interconnected multi-socket multicore processor nodes often with attached accelerator devices. They also need effective tools and methods to track and assess their codes’ execution performance as they aim to get ready for production on current or prospective exascale computer systems. This tutorial presents the methodology developed and applied over several years within the EU/EuroHPC Centre of Excellence Performance Optimisation and Productivity (POP). Its focus is the hierarchy of execution efficiency and scaling metrics that identify the most critical issues and quantify potential benefits of remedies. The metrics can be readily compared and determined by a variety of tools for applications in any language employing standard MPI, OpenMP/OpenACC and other multi-threading and offload paradigms. Using their own notebook computers tutorial participants will follow exercises using widely-deployed open-source tools and provided performance measurements of actual HPC application executions (ranging from CFD to neuroscience), preparing them to locate and diagnose efficiency and scalability issues in their own parallel application codes.

RationaleThis workshop is designed to provide participants with a comprehensive understanding of molecular dynamics simulations, the use of the Gromacs software, and its application in simulating biomembranes. Participants will gain know-how in setting up and running simulations, analyzing results, and interpreting biomembrane dynamics.SyllabusGROMACS OverviewIntroduction to molecular dynamics (MD)Force fields in MDMD Simulations with GromacsMD TrajectoriesAdvanced Topics in MD SimulationsBiomembrane structure and compositionCase studies and research Applications

Python is increasingly used in high-performance computing projects. It can be used either as a high-level interface to existing HPC applications and libraries, as embedded interpreter, or directly.This course combines lectures and hands-on sessions. We will show how Python can be used on parallel architectures and how to optimize critical parts of the kernel using various tools.The following topics will be covered:Interactive parallel programming with IPythonProfiling and optimizationHigh-performance NumPyJust-in-time compilation with numbaDistributed-memory parallel programming with Python and MPIBindings to other programming languages and HPC librariesInterfaces to GPUsThis course is aimed at scientists who wish to explore the productivity gains made possible by Python for HPC.

AnnotationThis training is designed to equip participants with essential knowledge and skills to utilize the IT4Innovations (IT4I) supercomputing infrastructure effectively. Covering a spectrum of topics ranging from HPC system architecture to performance analysis basics, participants will gain insights into accessing, utilizing, and optimizing their computational workflows on high-performance clusters. Experienced instructors will facilitate each session, offering theoretical insights and hands-on exercises. Participants are encouraged to engage actively to maximize their learning experience. Registered participants will receive access to the HPC infrastructure.Target Audience and Purpose of the Course:This course is primarily crafted for application users with prior experience on smaller platforms, such as desktops, who now seek to leverage HPC capabilities to address problems surpassing the capabilities of their current resources. These users, often scientists from diverse fields, may not be developers of the applications they employ but require a deeper understanding of the underlying processes, particularly parallelization.The course aims to offer participants an introduction to the key aspects of HPC, with a specific focus on the practices at IT4I, addressing their pertinent queries along the way. LevelbeginnerLanguageEnglishAgenda1. HPC systems architecture and IT4I architecture introduction Brief introduction to the High-Performance Computing (HPC) in the world and in the Czech Republic. Presenting hardware architecture of the contemporary HPC systems, and basics every HPC user should know, such as scalability, top500, path to exascale, or EPI. 2. Accessing and using IT4I clustersFirst login.How to get your data to the cluster.How to log in to the cluster and prepare a computation environment.How to submit computational jobs. 3. Code development on the clusterRemote development on the cluster.Compiling and building software tools and custom code. 4. Parallel Programming Basics This section briefly describes how to run parallel applications that use threads and MPI. And, how to set the environment variables for a hybrid run to utilize the provided hardware efficiently. 5. Leveraging accelerated hardware (GPUs)In the GPU section of the course, attendees will understand the place of today’s GPUs in HPC. Their architecture will be introduced and compared with CPUs. Several GPU programming models will be briefly introduced, focusing on CUDA, and including a few hands-on exercises. 6. Performance Analysis BasicsWhy should we care about performance? This section will answer the question, introduce you to the basic concepts of a performance analysis, provide an overview of performance tools, and show how to start using them. TutorsFilip Vaverka holds a PhD degree in Computer Science from Brno University of Technology, Czech Republic. He is currently a member of the Infrastructure Research Lab at IT4Innovations National Supercomputing Center. His research focuses on highly scalable pseudo-spectral methods, accelerated and heterogeneous parallel computing. Ondřej Meca holds a PhD degree in Computer Science from VSB—Technical University of Ostrava, Czech Republic. He is currently a member of the Infrastructure Research Lab at IT4Innovations National Supercomputing Center. His research interests include verifying parallel algorithms, developing pre/post-processing algorithms for large-scale engineering problems, and developing highly scalable linear solvers. Jakub Homola is a PhD. student of Computational Science and a Research Assistant at IT4Innovations, VSB-TUO. He graduated in Computational and Applied Mathematics at VSB - Technical University of Ostrava, specializing in Computational Methods and HPC. His research and professional interests are FETI methods and GPU programming in CUDA, HIP and SYCL. He is partially involved in the EUPEX project, where he is working with the ESPRESO library to optimize it for future supercomputers with European processors. Radim Vavřík is a researcher in the Infrastructure Research Lab at IT4Innovations. He is interested in parallel computing, scalable algorithms design, GPU acceleration, code optimization, and performance analysis. He worked on a hydrological and flood modelling software, a high-performance heterogeneous platform for energy-efficient computing, a GPU acceleration of a highly parallel FEM library, and as a performance analyst in several EU CoEs and projects, including POP2, POP3, SPACE, and SCALABLE.

The CSC Summer School brings together students and researchers in different disciplines of scientific computing from all over the world. The school consist of lectures and hands-on training on parallel programming, code performance, and other necessary skills for the development of scientific software for supercomputers.General informationThe school is aimed for graduate students working in various fields of science, e.g., chemistry, physics, life sciences, engineering, or economics as well as industrial R&D professionals. Also undergraduates as well as post-docs will find the school very useful.The ten days of school are filled with coding, sports, leisure activities, and having good time together. It will be an experience you will never forget with the midnight sun in the Nordic night next to breathtaking wilderness!The following topics will be covered: Programming at supercomputer scale using message-passing interface (MPI)Using threading (OpenMP) for programming CPU nodes Programming GPUs with HIP/CUDA and OpenMPHybrid parallel programming combining MPI with OpenMP and HIP/CUDA (multi-GPU programming)Principles of efficient parallel input/output in supercomputersApplying learned concepts in HPC and AI applicationsThe participants will get a certificate containing a recommendation to your home university to grant you 5 ECTS credits for the school.You may also want to read CSC’s training stories article with positive experiences from 2024 or 2023.Learning outcomeThe goal is to build HPC expertise up to an intermediate/semi-advanced level during the school.After the summer school participants will be able to:Identify main parallel programming conceptsCreate parallel MPI programs with C, C++, or Fortran programming languagesCreate GPU-accelerated programsRecognize advanced parallel programming constructsRead and modify scientific applications implemented in C, C++, or Fortran and utilizing MPI, OpenMP, and/or HIP/CUDAPrerequisitesBasic skills to operate in a Linux environmentBasic experience in programming with C, C++, and/or Fortran, and to be familiar with their elementary concepts (e.g. variables, statements, control structures, functions, subroutines, modules/classes)Please see this link for a detailed description of expected background knowledge as well as for refresher material to help you to acquire the required knowledge before the school. Feel free to contact the course organizers if you have any questions.AccommodationAccommodation takes place in shared double rooms. A limited number of single person rooms is also available. Both room types have an en-suite shower and toilet. Linen and towels will be provided. ScheduleThere will be a bus transportation from downtown Helsinki to Nuuksio and back.Departure from Helsinki: 24.6.2025 at 9:00Arrival to Helsinki: 3.7.2025 by 15:00FeeThe early bird registration fee for all ten days is 2099€ in shared double rooms.Late registration price starting from 15.3. onwards is 2299€A single person room costs 270€ extra.Prices are with VAT included (25.5 %)The fee is all-inclusive. It includes accommodation, four meals and two coffee breaks on most days, social events, sport activities, refreshments, electronic course materials, and transportation from downtown Helsinki and back.Selection processThe participants will be selected based on their background information. Notification about acceptance will be sent within five business days after registration. Fee will be charged only upon selection.Deadline for early bird registration is 14.03.2025.Deadline for late registrations is 01.05.2025.Contact tuomas.rossi@csc.fi for course details or event-support@csc.fi for registration related questions.

This course will take place as an online event. The link to the online platform will be provided to the accepted registrants only.Fancy using High Performance Computing machines for AI? Fancy learning how to run your code one of Europe's fastest computers JUWELS Booster at FZJ?In this workshop, we will guide you through the first steps of using the supercomputer machines for your own AI application. This workshop should be tailored to your needs - and our team will guide you through questions like:How do I get access to the machines? How do I use the pre-installed, optimized software?How can I run my own code?How can I store data so I can access it fast in training?How can parallelize my training and use more than one GPU?In this workshop, we will try to get your code and your workflow running and would like to make the start on a supercomputer as smooth as possible. After this course, you are not only ready to use not only HAICORE but you have made your first step into unlocking compute resources even on the largest scale with a compute time application at the Gauss Supercomputing Center.This workshop will be held in a small group size with enough space to address your questions. Please give us an indication on what topics you are interested in and we will try to adjust.

The focus of this intermediate C++ course are the essential C++ software development principles, concepts, idioms, and best practices, which enable programmers to create professional, high-quality code. The course will give insight into the different aspects of C++ (object-oriented programming, functional programming, generic programming) and will teach guidelines to develop mature, robust, maintainable, and efficient C++ code.

This course will take place as an online event. The link to the streaming platform will be provided to the registrants only.GPU-accelerated computing drives current scientific research. Writing fast numeric algorithms for GPUs offers high application performance by offloading compute-intensive portions of the code to a GPU.The course will cover aspects of GPU architectures and programming. Focus is on the usage of the parallel programming language CUDA C++, which allows maximum control of NVIDIA GPU hardware. Examples of increasing complexity are used to demonstrate optimization and tuning of scientific applications.The foundations of GPU programming are covered in a dedicated Basic Course which include an introduction to GPU/parallel computing, programming with CUDA, GPU libraries, tools for debugging and profiling, and performance optimizations. Please see the CUDA Basics course for registration.This advanced course consists of modules providing more in-depth coverage of multi-GPU programming, modern CUDA concepts, CUDA Fortran, and portable programming models such as OpenACC and C++ parallel STL algorithms. The advanced modules will be taught from 7-11 July 2025.Note: The GPU Programming with CUDA course is held in two parts. This is the second course, which contains individual modules to allow for fine-grained selection of topics of relevance.Contents of Part 1: Basics of GPU Programming with CUDADate: 31. March - 4 April 2025, on-site at JSC (see separate announcement)The agenda of the basic course are given here for completeness. For registration, please see dedicated website.A) Introduction to GPUs and GPU ComputingB) Programming Model CUDAC) Tools for Debugging and ProfilingD) GPU Libraries (like cuBLAS, cuFFT)E) Introduction to Multi-GPU ProgrammingContents of Part 2: Advanced GPU ProgrammingDate: 7-11 July 2025 (this announcement)A) Advanced Multi-GPU Programming with MPIB) Advanced Multi-GPU Programming with NCCL and NVSHMEMC) Advanced and Modern CUDA Concepts (Cooperative Groups, CUB Primitives, Modern C++ Programming)D) CUDA FortranE) GPU Programming with Abstractions (OpenACC, Standard Language Programming (pSTL))Attendees are invited to pick and choose the parts of the advanced course (A - E) they want to attend. The advanced modules are mostly freestanding. Participants either need to attend the basics course or prove equivalent knowledge of GPU programming in order to participate in the advanced course.Programmers interested primarily in OpenACC may skip parts D and E of Basics of GPU Programming with CUDA and still choose part E from Advanced GPU Programming. Participation in the full Basics of GPU Programming with CUDA course, however, is recommended.

This course will take place as an on-site and in-person event. It is not possible to attend online.Contents:An introduction to the parallel programming of supercomputers is given. The focus is on the usage of the Message Passing Interface (MPI), the most often used programming model for systems with distributed memory. Furthermore, OpenMP will be presented, which is often used on shared-memory architectures.The first four days of the course consist of lectures and short exercises. A fifth day is devoted to demonstrating the use of MPI and OpenMP in a larger context. To this end, starting from a simple but representative serial algorithm, a parallel version will be designed and implemented using the techniques presented in the course.Topics covered:Fundamentals of Parallel ComputingHPC system architecturesshared and distributed memory conceptsOpenMPbasicsparallel constructdata sharingloop work sharingtask work sharingMPIbasicspoint-to-point communicationcollective communicationblocking and non-blocking communicationdata typesI/OcommunicatorsHybrid programmingToolsEvent page can be found here.Agenda The preliminary agenda can be found here. Prerequisites:Knowledge of either C, C++, Python, or Fortran, basic knowledge of UNIX/Linux (incl. command line, Linux shell) and a UNIX standard editor (e.g. vi, emacs)

The Train the Trainer Program is provided in conjunction with the regular course Parallel Programming Workshop (MPI, OpenMP and Advanced Topics). Whereas the regular course teaches parallel programming, this program is designed to train future trainers in parallel programming.Too few people can offer parallel programming courses at the level needed by scientists and PhD students who want to learn how to parallelize a sequential application or to improve parallel applications. Within Europe, only a few HPC centres provide such courses at European or national level. We would like to support additional trainers and centres to also deliver such courses throughout Europe or at least in their own countries.This course is a joint training event of HLRS and EuroCC@GCS, the German National Competence Centres for High-Performance Computing.

Distributed memory parallelization with the Message Passing Interface MPI (Mon, for beginners):On clusters and distributed memory architectures, parallel programming with the Message Passing Interface (MPI) is the dominating programming model. The course gives an introduction to basic features available since MPI-1. Hands-on sessions (in C, Fortran, and Python) will allow users to immediately test and understand the basic constructs of the Message Passing Interface (MPI).Shared memory parallelization with OpenMP (Tue, for beginners):The focus is on shared memory parallelization with OpenMP, the key concept on hyper-threading, dual-core, multi-core, shared memory, and ccNUMA platforms. This course teaches shared memory OpenMP parallelization. Hands-on sessions (in C and Fortran) will allow users to immediately test and understand the directives and other interfaces of OpenMP. Race-condition debugging tools are also presented.Intermediate and advanced topics in parallel programming (Wed-Fri):Topics are advanced usage of communicators and virtual topologies, one-sided communication, derived datatypes, MPI-2 parallel file I/O. MPI-3.0 introduced a new shared memory programming interface, which can be combined with MPI message passing and remote memory access on the cluster interconnect. It can be used for direct neighbor accesses similar to OpenMP or for direct halo copies, and enables new hybrid programming models. Several aspects of hybrid mixed model MPI+OpenMP parallelization are discussed in the MPI-4.1 and OpenMP-4.5(5.0) advanced topics.

ESiWACE3 CoE - Simulation of Weather and Climate in EuropeDate & LocationDates: 21–23 October 2025Location: Jülich Supercomputing Centre (JSC), Jülich, GermanyCosts: There is no participation fee. ESiWACE3 will cover all costs related to the event, including catering during the hackathon and the hackathon dinner. However, participants are responsible for their own travel and accommodation expenses.ModalityHybrid: In-person attendance is strongly encouraged to maximize collaboration and hands-on support. Online participation is an option only if on-site attendance is not possible. However, at least one member from each team must attend in person. Other team members may participate remotely if needed.Scope and ObjectivesThis hackathon is designed for researchers and developers working with Earth System Models (ESMs) who are looking to prepare their applications for the first EuroHPC exascale supercomputer, JUPITER. The event provides an opportunity to work directly on porting and optimizing climate and weather models for this next-generation HPC system. You will have the chance to pick up advice and tips from HPC experts and fellow model developers in a friendly and supportive environment. This hackathon is a great opportunity to collaborate, troubleshoot, and refine your code with expert guidance. You will also work on addressing challenges in performance and portability, helping to prepare your Earth System Models for JUPITER.JUPITER will be the first exascale system in Europe, featuring NVIDIA GH200 Grace Hopper Superchips. This event will provide hands-on experience in adapting ESMs to fully exploit JUPITER’s architecture.Who Can Apply?The hackathon is open to researchers affiliated with institutions in EuroHPC Joint Undertaking countries or Inclusiveness Target Countries. We encourage applications from teams of two to four members, with a mix of expertise in Earth system modeling and HPC performance tuning.Teams should propose well-defined projects that can be realistically tackled within the three-day event. While expert HPC support will be available, participants should have a solid understanding of their codebase to effectively engage in the porting and optimization process. Preference will be given to open-source applications, and we strongly encourage applications from early-career scientists and researchers from under-represented groups.Application and RegistrationInterested teams must submit a short proposal outlining their project, including a brief description of the code or model, the specific challenges or goals for JUPITER adaptation, and the expected outcomes of participation in the hackathon.Register through the online form available at: https://forms.office.com/r/RiC1BiCZRbPrizeTo reward the winning team and support further collaboration, the HPC services provider company Do It Now will sponsor €1500 to cover the expenses of attending an academic conference within Europe. This funding will help the winning team present their work and engage in valuable discussions at a relevant conference in Europe.Important DatesApplication deadline: 30 June 2025Notification of acceptance: 15 July 2025ContactFor any questions regarding the event or for an informal chat about whether your idea might be suitable, feel free to reach out to the organizers at: l.zou@fz-juelich.de or l.hoffmann@fz-juelich.deWe look forward to your participation in this exciting opportunity to prepare Earth system models for the exascale era!

This course is targeted at researchers with basic knowledge in numerical simulation, who would like to learn how to visualize their simulation results on the desktop but also in Augmented Reality and Virtual Environments. The two-day workshop gives a short overview over scientific visualization in general, followed by a hands-on introduction to 3D desktop visualization with VISTLE (webpage on the EXCELLERAT P2 service portal) and COVISE. Participants will further learn how to build interactive 3D Models for Virtual Environments and how to set up an Augmented Reality visualization.

MPI is a communication protocol for parallel programming based on message exchange between individual processes. These processes can be executed on distributed memory systems with multiple nodes. This makes MPI scalable on systems larger than single computers. MPI offers a range of tools to maintain the flow of information between individual processes. This enables the execution of a parallel program that can be divided into several smaller parts. In the course of necessary communication, overhead always occurs, which normally limits the scalability of a parallel program. However, a properly optimized program opens up the possibility of using MPI on a distributed memory system (e.g., cluster or supercomputer) with satisfactory efficiency, where thousands or tens of thousands of nodes can be used. This course also offers the opportunity for intensive exchange with the instructors and other course participants.

This advanced C++ course provides both advanced C++ programming techniques and software design insight to help developers to create professional, high-quality code. The course provides deeper insight into C++ templates (type traits, SFINAE, C++20 concepts and forwarding references) and advice on how to create safer and cleaner function and class interfaces. Additionally, it provides an introduction to software design. For that purpose, it demonstrates the gravity of bad dependency management and shows how to properly reduce dependencies by means of design patterns.