From Fortran to Python: A Conversation Across Generations of Scientific Computing
summary
Retro account of punch‑card workflows and a recreated chemical ODE for EDC cracking that runs as a compiled Fortran subroutine compiled with f2py imported into Python, with repo and build notes (Meson/Ninja caveat). Prototype Python version used for orchestration and visualization; numerical core replaced by Fortran for benchmarks showing up to 700× speedup vs pure Python on many-step integrations. Emphasis on concrete workflow differences and tool interoperability.
AI-written summary · about
chapters (13)
- 0:00Intergenerational scientific computing
- 3:05Fortran’s role in science
- 5:57What is Fortran
- 7:02Fortran’s workflow in the 1980s and 1990s
- 8:32Fortran vs Python for numbers
- 9:54Chemical ODE Fortran vs Python
- 12:06Run Fortran in Python with f2py
- 13:57Performance Fortran vs Python
- 15:25Julia in scientific computing
- 16:32Key takeaways and tools
- 17:19Punch cards to modern coding
- 19:07AI assistants in science
- 20:35Q&A on punch cards
related talks
36mPython, C, C++, and Fortran Relationship Status: It’s Not That Complicated
06mLessons learned from using the C Foreign Function Interface
49mPython for High Performance and Scientific Computing
32mNot Invented Here: Porting Scientific Software to Python- 32mNot Invented Here: Porting Scientific Software to Python
33mBuilding with Multiple Languages: A Python-First Approach
44mPython's Bright Future in Science
30mEmpowered by Python - A success story