Interactive C++ for Data Science
summary
Shows a working interactive C++ stack: Cling as an incremental compiler/JIT that runs full C++ (templates, lambdas, virtual inheritance), xeus-cling for Jupyter notebooks, cppyy for Python bindings and interactive CUDA for GPU kernels. Demonstrations include reading ROOT files to probe objects from exabyte-scale datasets, live GPU simulation and visualization on clusters like Titan/Summit, template instantiation on demand (example: instantiate a C++ template from a C program in ~120 lines) and an automatic differentiation plugin that produces and runs derivatives.
AI-written summary · about
chapters (25)
- 0:00Intro
- 0:17Outline
- 3:05Acknowledgement & Disclaimer
- 4:04Introduction
- 4:43Choice of translation
- 5:55Exploratory Programming
- 7:34Interactive C++. Key Insights
- 10:08Interpreters for C++
- 11:10Interpreting C++. Cling
- 14:02Data Science
- 16:22Xeus-Cling. C++ in Notebooks
- 19:14Interactive CUDA C++
- 25:48ROOT - Scientific Data Analysis
- 27:11Dynamic Scopes. Runtime Lookup
- 29:22Impact of Interactive C++ in Physic
- 31:03Automatic Language Interop Classification of Prior Work
- 32:34Automatic Language InterOp. Python
- 35:16Automatic Language InterOp. D
- 36:17Automatic Language InterOp. Julia How does it work? - Runtime Features - Interactive CH REPL
- 39:41Compiler As A Service. Extensions
- 41:37Lifelong Optimization
- 44:24Incremental Compilation in LLVM
- 44:53Instantiating a C++ template in C
- 47:28Conclusion
- 48:23Thank You!
related talks
38mJumping into Jupyter Notebooks
36mMaking interactive visualizations easy, inside and outside of notebooks
38m"Emmy: Moldable Physics and Lispy Microworlds"
40mUse animated data stories in Jupyter to present
14mB2: Bridging Code and Interactive Visualization in Computational Notebooks
42mInteractive Data Analysis: Visualization and Beyond
11mJavaScript Visualized: A Script's Life
43mAdvanced Data Visualizations In Jupyter Notebooks