Transonic documentation

Transonic is a pure Python package (requiring Python >= 3.9) to easily accelerate modern Python-Numpy code with different accelerators (currently [Cython], [Pythran], [Numba] and [JAX], but potentially later [Cupy], [PyTorch], [Weld], [Pyccel], etc…).

The accelerators are not hard dependencies of Transonic: Python codes using Transonic run fine without any accelerators installed (of course without speedup)!

Get started

• Overview
  • A short tour of Transonic public API
  • Make the Pythran/Cython/Numba files and compile the extensions
• Install and configure
• Supported backends
  • Unified code for different backends
  • Pythran backend
  • Cython backend
  • Numba backend
• Packaging when using Transonic

Examples

• Nearly as with Pythran
  • Function calls
• Accelerated functions with type hints
• Cached Just-In-Time compilation
• Define accelerated blocks
• Object oriented programming!
• Ipython and the “compile_at_import” mode
  • Summary
• IPython + jit decorator
• inline functions
• Writing benchmarks
  • Unified code for different backends
  • Benchmark rotation vector
  • Benchmark projection vector
  • Benchmark expression broadcast
  • Comparison Numba vs Pythran (JIT)
  • Ahead-of-time compilation
• Benchmark in Jupyter with the jit decorator

API Reference

Here is presented the organization of the package and the documentation of the modules, classes and functions.

transonic.aheadoftime	User runtime API for the ahead-of-time compilation
transonic.analyses	Code analyses
transonic.backends	Backends for different accelerators
transonic.compiler	Use Pythran to create extensions
transonic.config	Configuration
transonic.dist	Utilities for the setup.py files
transonic.justintime	Cached JIT compilation
transonic.log	Logging
transonic.mpi	Minimalist MPI module
transonic.run	Command line transonic
transonic.signatures	Compute function signatures
transonic.typing	Create Pythran signatures from type hints
transonic.util	Internal utilities

More

• Transonic forge on Heptapod
• Transonic in PyPI
• Release notes
• Roadmap
• Thanks
• How to contribute to Transonic
• Advice for FluidDyn developers
• For developers

Indices and tables

• Index

• Module Index

• Search Page