Building Fast Packages Faster: Julia as a Backend to Python and R

on YouTube ↗34m

" summary

Concrete claim: Python and R wrappers diffeqpy and diffeqr bridge to DifferentialEquations.jl from SciML, install Julia and GPU stacks automatically, and run Julia solvers from Python/R. Measured results show Julia-based GPU solvers give 20x–100x acceleration versus PyTorch and Jax on benchmark problems. The packages translate problem definitions into Julia types, JIT-compile optimized fused kernels, and the demos report successful speedups and functioning GPU workflows.

AI-written summary · about

" chapters (10)

1. 0:00Introduction
2. 1:04What is Julia
3. 3:54Julia’s community
4. 7:50SciML ecosystem
5. 14:57Differential equations in Python/R
6. 22:40What is SciML
7. 27:01Julia with PyTorch
8. 28:27Adoption challenges
9. 31:51Two-language problem
10. 33:36Q&A

" related talks

• 36mAccelerating Differential Equations In R and PythonChris Rackauckas · PyData Global 2020
• 32mJulia for PythonistasKelly Shen · PyData New York City 2019
• 43mJulia for Data Analysis features, interfaces and future directionsSimon Byrne · PyData London 2016
• 01h 00mPython and Julia: the best of friends?Malcolm Sherrington · PyData London 2016
• 01h 27mIntroduction to Julia for Scientific Computing and Data ScienceDavid Higgins, Robert Schwarz · PyData Berlin 2017
• 01h 37mTech Tutorial: Use Julia for your next scientific computing job!P. Szufel · International Data Science Conference 2021
• 01h 07mJulia TutorialChase Coleman · PyData DC 2016
• 01h 57mIntroduction to Julia for PythonistasJohn Pearson · PyData Carolinas 2016