Best SW Practices#

Good software practices are vital for successful scientific computing. This page outlines the best methods to follow when developing and running software on HPC systems.

Coding Standards#

Adhering to coding standards ensures your code is readable, maintainable, and less prone to errors.

  • Follow a Style Guide: Include a style guide relevant to your programming language.

  • Use Meaningful Names: Name variables and functions descriptively.

  • Comment Wisely: Write comments that explain the ‘why,’ not the ‘what.’

Version Control#

Utilizing a version control system like Git helps track changes, allows collaboration with others, and is essential for reputability.

  • Commit Regularly: Make frequent, small commits.

  • Use Meaningful Commit Messages: Describe what and why changes were made.

  • Keep a Changelog: Document significant changes in a human-readable file.

Documentation#

Proper documentation helps others (and you in the future) understand your code.

  • Write inline comments: Explain complex or non-intuitive parts of the code.

  • Maintain a README: Include setup, usage, and other essential information.

  • Use automated documentation tools: Tools like Doxygen or Sphinx can help.

Testing#

Implementing automated tests ensures that your code behaves as expected.

  • Write Unit Tests on individual code units for correctness.

  • Perform Integration Tests between different code parts.

  • Automate testing using Continuous Integration tools to run tests automatically.

Dependency Management#

Managing dependencies effectively avoids conflicts and ensures reputability.

  • Use Dependency Management tools like Conda or virtualenv.

  • Specify exact versions: Note the specific versions of dependencies.

Performance Considerations#

Consider performance throughout the development process.

  • **Profile often: Identify bottlenecks and areas for optimization.

  • **Consider parallelization: Utilize parallel computing resources.

Security#

Please follow the best practices to keep your code and data secure.

  • Keep Sensitive Information Private: Use environment variables for secrets.

  • Stay Updated: Keep dependencies updated to avoid known vulnerabilities.

Collaboration#

Effective collaboration fosters quality and productivity.

  • Communicate Openly: Keep lines of communication open with your team.

  • Use Collaboration Platforms: GitHub or GitLab facilitate collaboration.

These best practices will help develop high-quality, maintainable, and performing software suitable for HPC and other computational environments. Please always seek feedback from peers, and don’t hesitate to consult the HPC support team for specific guidance.