An Introduction to Scientific Software Tools & Parallel Algorithms (SSPA)

An Introduction to Scientific Software Tools & Parallel Algorithms (SSPA)#

Item

Link

Jupyter Book Status

Jupyter Book Status

Rendered book

https://luca-heltai.github.io/sspa/

Course repository

luca-heltai/sspa

Author

Luca Heltai

Course materials for a 30-hour PhD-level class (10 × 3h sessions) on practical tools for scientific software and introductory parallel algorithms.

Overview#

This course provides a comprehensive, hands-on introduction to the tools and workflows used in modern scientific software development and an applied introduction to parallel algorithms and performance analysis. Topics include:

  • Unix shell and HPC usage (ssh, tmux, Slurm job scheduler)

  • Version control using Git (local and collaborative workflows)

  • Automated documentation (Doxygen for C/C++, Sphinx/MyST for Python and general docs)

  • Unit and functional testing (Google Test for C++, pytest for Python)

  • Containerization and reproducibility (Docker, Apptainer/Singularity)

  • Continuous Integration and automation (GitHub Actions)

  • Basics of parallel computing: speedup, efficiency, Amdahl’s and Gustafson’s laws

Learning outcomes#

By the end of the course, students will be able to:

  • Operate comfortably in a Linux command-line environment and write basic Bash scripts.

  • Submit and monitor jobs on a Slurm-based cluster (we provide a Dockerized simulator for exercises).

  • Use Git to manage projects, branch, merge, resolve conflicts, and collaborate via pull requests.

  • Produce documentation with Doxygen and Sphinx/MyST, including API references and tutorial pages.

  • Write unit and functional tests for C++ and Python code and integrate them into CI.

  • Build and run Docker and Apptainer containers to package software environments for reproducibility.

  • Set up simple CI workflows (GitHub Actions) to run tests and build artifacts automatically.

  • Measure simple parallel programs, compute speedup and efficiency, and interpret results using Amdahl’s and Gustafson’s laws.

Quick start#

  1. Browse the compact course book in jupyterbook/ (index and lectures/).

  2. To build the book locally (requires jupyter-book):

pip install -U jupyter-book
jupyter-book build jupyterbook

  1. Hands-on examples commonly use Docker. See docker-images/ for Dockerfiles and usage notes if present.

License#

Content: CC-BY. Code examples: MIT. See LICENSE for details.

Course repository structure#

Typical layout (some folders may be added later):

  • jupyterbook/ — the course book (index, TOC, lectures)

  • lectures/ or jupyterbook/lectures/ — session notes and lab instructions

  • code/ — example programs used in lectures

  • exercises/ — assignment descriptions and starter files

  • docker-images/ — Dockerfiles and compose files for the mini-cluster

  • .github/workflows/ — example CI pipelines used to build/test the material

See jupyterbook/index.md for a curated table of contents and links to each lecture page.

Use the table of contents to navigate the 10 sessions (each ~3h). See the top-level repository README for quick setup instructions.

Below is a quick summary of the course topics collected in this Jupyter Book:

  • Unix shell and HPC environments (Slurm)

  • Version control with Git

  • Documentation (Doxygen, Sphinx)

  • Testing (gtest, pytest)

  • Containerization (Docker, Apptainer)

  • Continuous Integration (GitHub Actions)

  • Parallel algorithms and performance analysis

Use the table of contents to navigate the 10 sessions (each ~3h).

Contents