Repository files navigation

cuDF (pronounced "KOO-dee-eff") is a GPU DataFrame libraryfor loading, joining, aggregating, filtering, and otherwisemanipulating data. cuDF leverageslibcudf, ablazing-fast C++/CUDA dataframe library and theApacheArrow columnar format to provide aGPU-accelerated pandas API.

You can importcudf directly and use it likepandas:

importcudftips_df=cudf.read_csv("https://github.com/plotly/datasets/raw/master/tips.csv")tips_df["tip_percentage"]=tips_df["tip"]/tips_df["total_bill"]*100# display average tip by dining party sizeprint(tips_df.groupby("size").tip_percentage.mean())

Or, you can use cuDF as a no-code-change accelerator for pandas, usingcudf.pandas.cudf.pandas supports 100% of the pandas API, utilizing cuDF forsupported operations and falling back to pandas when needed:

%load_extcudf.pandas# pandas operations now use the GPU!importpandasaspdtips_df=pd.read_csv("https://github.com/plotly/datasets/raw/master/tips.csv")tips_df["tip_percentage"]=tips_df["tip"]/tips_df["total_bill"]*100# display average tip by dining party sizeprint(tips_df.groupby("size").tip_percentage.mean())

• Try cudf.pandas now: Explorecudf.pandas on a free GPU enabled instance on Google Colab!
• Install: Instructions for installing cuDF and otherRAPIDS libraries.
• cudf (Python) documentation
• libcudf (C++/CUDA) documentation
• RAPIDS Community: Get help, contribute, and collaborate.

See theRAPIDS install page forthe most up-to-date information and commands for installing cuDFand other RAPIDS packages.

• CUDA 11.2+
• NVIDIA driver 450.80.02+
• Volta architecture or better (Compute Capability >=7.0)

cuDF can be installed viapip from the NVIDIA Python Package Index.Be sure to select the appropriate cuDF package dependingon the major version of CUDA available in your environment:

For CUDA 11.x:

pip install --extra-index-url=https://pypi.nvidia.com cudf-cu11

For CUDA 12.x:

pip install --extra-index-url=https://pypi.nvidia.com cudf-cu12

cuDF can be installed with conda (viaminiforge) from therapidsai channel:

conda install -c rapidsai -c conda-forge -c nvidia \    cudf=25.04 python=3.12 cuda-version=12.8

We also providenightly Conda packages built from the HEADof our latest development branch.

Note: cuDF is supported only on Linux, and with Python versions 3.10 and later.

See theRAPIDS installation guide for more OS and version info.

See buildinstructions.

Please see ourguide for contributing to cuDF.