class: center, middle # SymPy Code Generation Tutorial http://sympy.org/scipy-2017-codegen-tutorial ### SciPy 2017 ### 07/10/2017, 8:00 AM --- ## Who are we? .left-column75[ ### Aaron Meurer **@asmeurer** - Researcher ERGS, University of South Carolina - Lead developer of SymPy ] .right-column25[ <img src="asmeurer.jpeg" alt="asmeurer" style="width:200px"> ] --- ## Who are we? .left-column75[ ### Jason Moore **@moorepants** - Faculty, Mechanical and Aerospace Engineering, University of California, Davis - SymPy and PyDy Developer ] .right-column25[ <img src="moorepants.jpeg" alt="moorepants" style="width:200px"> ] --- ## Who are we? .left-column75[ ### Björn Dahlgren **@bjodah** - PhD student, KTH Royal Institute of Technology (Stockholm, Sweden) - SymPy contributor ] .right-column25[ <img src="bjodah.jpeg" alt="bjodah" style="width:200px"> ] --- ## Who are we? .left-column75[ ### Kenny Lyons **@ixjlyons** - PhD student, University of California, Davis ] .right-column25[ <img src="ixjlyons.jpg" alt="ixjlyons" style="width:200px"> ] --- ## This tutorial ### We will learn about code generation with SymPy ### Basic knowledge of the SymPy library is assumed --- class: center, middle ## SymPy Python library for symbolic mathematics. Version 1.1 released last Thursday (July 6, 2017). Make sure you have it! <img src="sympy-notext.svg" alt="sympy logo" style="width:200px"> --- ## What is code generation? -- For the purposes of this tutorial, "code generation" (or just "codegen") refers to the automatic translation of a symbolic SymPy expression into code for another language --- ## What is code generation? For example: .left-column[ *SymPy expression* ```py abs(sin(pi*x)) ``` ] .right-column[ *C code* ```C fabs(sin(M_PI*x)) ``` ] --- ### Some of the languages supported - C - C++ - Fortran - MATLAB/Octave - Python (NumPy/SciPy) - Julia - Mathematica - Javascript - LLVM - Rust - Theano - Tensorflow - Easy to extend to others... --- background-image: url(codegen1.png) ### What is code generation -- background-image: url(codegen2.png) --- background-image: url(abstraction.png) --- ### Why do code generation? -- 1. SymPy can deal with mathematical expressions in a high-level way. For example, it can take symbolic derivatives. -- 2. Using code generation avoids mistakes that come from translating mathematics into low level code. -- 3. It's possible to deal with expressions that are otherwise too large to write by hand. -- 4. Some "mathematical" optimizations are possible, which a normal compiler would not be able to do. --- ### Outline for today - Intro to SymPy Expressions [30 minutes] - Code Printers [30 minutes] - The Easy Way: High Level Generation (lambdify, ufuncify) [30 minutes] - The Harder Way: Code generation and compilation [1 hour] - Cythonizing Your Code (manually and autowrap) [30 minutes] - Bonus materials (if there is time) --- class: center, middle This tutorial is intended to be interactive. Please stop us to ask questions at any time. --- class: center, middle # Let's begin!