Resources

Welcome!

This page is for collecting and compiling resources about computational skills and best practices for the group.

Ideas for topics to include:

Fundamental computational skills

• navigating a terminal
  • basic bash skills
  • scripting
  • things to never ever do bc you could delete your entire OS
• using IDEs
• debugging
• standards and conventions (pep8, etc)
• conda & environment management
• version control: git
  • Branch/fork workflows
  • GitHub/GitLab/Bitbucket
  • Atlassian tools: JIRA, Confluence, Bitbucket & GitHub equivalents
• Software development life cycle/workflows
  • Task management/Kanban/Scrum/Agile-ish concepts
• pull requests and collaborative code contributing
  • "issues"/tickets and how to use those in branch/fork workflows
  • usage for things like QA documentation practices (NQA-1)
  • Code reviews: like "peer review" process of papers, but for software
• Documentation: why you need to write it & how to write GOOD documentation
  • No, the code should not document itself (as your sole form of documentation)
  • Sphinx, Doxygen, etc
• software design concepts
  • Scoping work
  • Technical requirements & requirements management
    • Configuration management
  • QA requirements, risks, failure modes
  • V&V
  • automated testing

• writing clean reusable code

• test-driven development

  • fail fast, fail often

• is this the right tool for the problem at hand?

  • how do you assess that?

• parallelization: when it's good, when it's not

• visualization tools (ParaView, etc)

Intermediate to advanced skills

• Object-oriented programming
• CI/CD
• GPU vs MPI
• CUDA 101
• CMake

• Make
• Dependency management
• Using open source libraries
• how to choose a language based on your applications
• data management
• memory management
• precautions re: using pre-existing tools
  • and writing code that "works" but is wrong
  • or writing code that makes "pretty" pictures but is useless
• design patterns & antipatterns
• Using IDEs to debug on clusters with ssh/scp connections
• Debugging with GPUs
• single vs multidevice GPU
  • precautions
  • when it becomes slower than multithreaded CPUs (data transfers/IO)
• profiles and formatting tools/linters
• Meshing (cubit, etc)
• minimum viable product, "rapid prototyping"

Languages and code skills

• C++
• Python
• FORTRAN
• CUDA
• shell scripting
• API design
• OOP concepts in general
• Namespaces/scopes
• public/private/protected