cudf Developer Guide

Note

This document assumes familiarity with the overall contributing guide.

cuDF is a GPU-accelerated, pandas-like DataFrame library. Under the hood, all of cuDF’s functionality relies on the CUDA-accelerated libcudf C++ library. Thus, cuDF’s internals are designed to efficiently and robustly map pandas APIs to libcudf functions. The goal of this guide is to provide specific guidance for cuDF Python developers. It covers the structure of the Python code and discusses best practices.

• Library Design
  • The Frame layer
  • The Column layer
  • Frame operations on Columns
  • Copy-on-write
• Contributing Guide
  • Directory structure and file naming
  • Code style
  • Deprecating and removing code
  • pandas compatibility
• Writing documentation
  • Docstrings
  • Published documentation
  • Comparing to pandas
  • Writing documentation pages
  • Building documentation
• Testing cuDF
  • Tooling
  • Test organization
  • Test contents
• Benchmarking cuDF
  • Benchmark organization
  • Running benchmarks
  • Benchmark contents
  • Comparing to pandas
  • Testing benchmarks
  • Profiling
  • Advanced Topics
• String UDF memory management
  • Numba memory management and the Numba Runtime (NRT)
  • Data structures
  • NRT Requirements
  • Implementation
  • String Lifecycle Details