際際滷

際際滷Share a Scribd company logo

Numba lightning

Download as KEY, PDF
Travis Oliphant
Travis Oliphant

Numba is a Python compiler that uses type information to generate optimized machine code from Python functions. It allows Python code to run as fast as natively compiled languages for numeric computation. The goal is to provide rapid iteration and development along with fast code execution. Numba works by compiling Python code to LLVM bitcode then to machine code using type information from NumPy. An example shows a sinc function being JIT compiled. Future work includes supporting more Python features like structures and objects.

1 of 11
Numba NumPy-aware dynamic Python compiler Lightning Talk Travis E. Oliphant PyData NYC October 27, 2012
Motivation Python is great for rapid development and high-level thinking-in-code It is slow for interior loops because lack of type information leads to a lot of indirection and extra code.
Motivation NumPy users have a lot of type information --- but only currently have one-size 鍖ts all pre-compiled, vectorized loops. Many new features envisioned will need the ability for high-level expressions to be compiled to machine code.
Goal: rapid iteration and development + fast code execution Numba should be the worlds best array-oriented compiler.
NumPy + Mamba = Numba Python Function Machine Code LLVM-PY LLVM Library ISPC OpenCL OpenMP CUDA CLANG Intel AMD Nvidia Apple ARM
Example Numba @jit(f8(f8)) def sinc(x): if x==0.0: return 1.0 else: return sin(x*pi)/(pi*x)
Status and Future My early bytecode branch further developed by Jon Riehl (Resilient Science) sponsored by Continuum Analytics, Inc. --- interprets bytecode directly Current trunk works with AST directly and making rapid progress - Mark Florrison (minivect) - Siu Kwan Lam (pymothoa)
Demo
RoadMap Numba 0.2 available now Github trunk has many changes -- 0.3 will support - structures - Python code - objects with inheritance Check out NumbaPro for cutting-edge features
Software Stack Future? Plateaus of Code re-use + DSLs SQL R TDPL Matlab Python OBJC C FORTRAN C++ LLVM
Join Us! http://numba.pydata.org https://store.continuum.io/cshop/numbapro

More Related Content

Numba lightning

  • 1. Numba NumPy-aware dynamic Python compiler Lightning Talk Travis E. Oliphant PyData NYC October 27, 2012
  • 2. Motivation Python is great for rapid development and high-level thinking-in-code It is slow for interior loops because lack of type information leads to a lot of indirection and extra code.
  • 3. Motivation NumPy users have a lot of type information --- but only currently have one-size 鍖ts all pre-compiled, vectorized loops. Many new features envisioned will need the ability for high-level expressions to be compiled to machine code.
  • 4. Goal: rapid iteration and development + fast code execution Numba should be the worlds best array-oriented compiler.
  • 5. NumPy + Mamba = Numba Python Function Machine Code LLVM-PY LLVM Library ISPC OpenCL OpenMP CUDA CLANG Intel AMD Nvidia Apple ARM
  • 6. Example Numba @jit(f8(f8)) def sinc(x): if x==0.0: return 1.0 else: return sin(x*pi)/(pi*x)
  • 7. Status and Future My early bytecode branch further developed by Jon Riehl (Resilient Science) sponsored by Continuum Analytics, Inc. --- interprets bytecode directly Current trunk works with AST directly and making rapid progress - Mark Florrison (minivect) - Siu Kwan Lam (pymothoa)
  • 9. RoadMap Numba 0.2 available now Github trunk has many changes -- 0.3 will support - structures - Python code - objects with inheritance Check out NumbaPro for cutting-edge features
  • 10. Software Stack Future? Plateaus of Code re-use + DSLs SQL R TDPL Matlab Python OBJC C FORTRAN C++ LLVM
  • 11. Join Us! http://numba.pydata.org https://store.continuum.io/cshop/numbapro

Editor's Notes

  1. \n
  2. \n
  3. \n
  4. \n
  5. \n
  6. \n
  7. \n
  8. \n
  9. \n
  10. \n
  11. \n