Tutorials

This intensive tutorial provides practical training in open science workflows for large-scale data analysis using datasets from CERN, NASA, ESA, Zenodo, and arXiv. Over two 4-hour sessions, participants will gain hands-on experience with enterprise-grade tools such as GitHub, GitHub Actions, Jupyter Notebooks, and cloud-based platforms. A distinctive feature of this tutorial is its integration of AI workflows, particularly Large Language Models (LLMs) in Retrieval-Augmented Generation (RAG), to enhance literature reviews, meta-analyses, and scientific writing. Each session includes guided exercises focused on automating workflows, analyzing data at scale, and applying ML pipelines in reproducible and collaborative research environments. Participants should bring a laptop with internet access and a browser, and have basic familiarity with Linux terminal and Python; Git basics are helpful but not required.

This tutorial addresses:
1/ Real-world problems in accessing and analysing big scientific data.
2/ Automation of complex workflows using GitHub Actions.
3/ Open Science best practices for reproducibility and collaborative publishing.
4/ Integration of cutting-edge LLMs for research assistance and meta-analyses.

The learning outcomes are as follows:

• Connect to and access open scientific datasets;
• Automate data pipelines using GitHub Actions;
• Analyse and visualise large datasets with Jupyter Notebooks;
• Apply ML models for classification and clustering tasks;
• Use LLMs in RAG workflows for literature reviews and meta-analyses;
• Optimise the use of free/low-cost computational resources for research.

Duration: 8 hours
Participants must register: https://indico.redclara.net/e/introhpccarla2025

HPC is based in a complex set of infrastructures, networking, software and data, with the ultimate objective of solve demanding and CPU-intensive tasks in the most efficient and fast way without loosing accuracy. In the very heart of any HPC system, lies the distributed programming model. The most important models nowadays are the shared-memory multicore paradigm, the distributed-memory cluster paradigm and the GPU based paradigm, each one with its own programming tools.

In this Tutorial, we will learn the basis of the first two models, OpenMP and MPI, that are the most representative tools for shared memory and distributed memory architectures respectively.

This Tutorial is totally hands-on, and is designed for beginners and students of the first courses of programming, with knowledge on C language. Students will interact with a real cluster, and will practice the very first -yet real- exercises in parallel architectures. The goal is to serve as a motivation for deeper understanding of HPC world, and as a basis for GPU-intensive architecture study.

Student’s prerequisites

• Tutorial will be in C and C++. People with experience in other programming languages are welcome too
• Simple understanding of how to use compilers
• Basic experience with Linux systems

Equipment
The tutorial is to be delivered on site and students will be invited to bring their own laptop. The laptop require access to internet and a Browser, and SSH connection using BitVise, SSH Client or similar.

This Tutorial is based on material from the instructors, that will be given in advance.

Robinson Rivas-Suarez

Gilberto Diaz

Luis Alejandro Torres Niño

This 4-hour tutorial introduces participants to the foundations of quantum computing and its simulation through the Q-Team environment. The first part covers the transition from classical to quantum computing, including logic gates, transistors, ALUs, and the principles of quantum information theory. Participants will explore qubits, quantum gates, and simulators to understand the building blocks of quantum algorithms. The second part focuses on hands-on exercises with the Q-Team simulator, where attendees will design and simulate algorithms using fundamental gates (Pauli X, Y, Z, Hadamard, phase shift, controlled gates) and circuits. By the end of the session, participants will be able to implement and document basic quantum algorithms within the simulator. The tutorial is intended for students in mechatronics, electronics, IT, and related areas, with prior knowledge of linear algebra, combinational logic, and complex analysis.

The tutorial topics include two modules.

• What the hell is quantum computing? High performance computing, cloud, and data centers. The computer. The microprocessor. The Arithmetic logic unit (ALU). Logic gates. Transistors and miniaturization. Quantum physics. Quantum information theory. Qubit. Qubit from superconductors. Quantum gates. Quantum processors. Quantum simulators. Q-Team.
• Simulating quantum algorithms with Q-Team. Qubit. Quantum Gates. Pauli X. Pauli Y. Pauli Z. Hadamard. Phase shift. Controlled gates. Quantum circuits. Quantum computing from classical computing. Most Common Algorithms.

This tutorial offers a practical and conceptual introduction to distributed deep learning, an essential approach for scaling large AI models across multiple GPUs and compute nodes. Participants will learn about data, model, and pipeline parallelism, and explore state-of-the-art frameworks such as PyTorch Distributed, DeepSpeed, and Ray. The program combines conceptual sessions on distributed training strategies and system design with a hands-on lab using Google Colab, where attendees will set up distributed environments and implement example workflows. The tutorial also covers performance tuning, communication overhead, and scalability considerations, preparing participants to train modern AI models efficiently on HPC or cloud platforms. It is intended for participants with intermediate knowledge of machine learning and deep learning fundamentals, familiarity with Python, and access to a laptop capable of running Google Colab (no local software installation required).

By the end of this tutorial you’ll:
1/ understand the core concepts behind distributed deep learning
2/ Learn about different distributed and parallel strategies and when to use them;
3/ Get hands-on experience setting up and running distributed training using common frameworks.

The tutorial timeline includes:
1/ Introduction to Distributed Deep Learning and Motivation. Data versus Model Parallelism.
2/ TensorFlow Distributed Tools PyTorch Distributed Tools DeepSpeed Ray.
3/ Hands on Tutorial on Adding Distributed Tools with PyTorch, DeepSpeed and Ray. (Jupyter/Colab Notebook).