Dask support for distributed GDF object
Switch branches/tags
Nothing to show
Clone or download
kkraus14 Merge pull request #47 from rapidsai/fea-conda-builds
[REVIEW] Enable conda builds and uploads
Latest commit 8faedae Nov 7, 2018

README.md

Dask GPU Dataframes

A partitioned gpu-backed dataframe, using Dask.

Setup from source

Setup from source repo:

  1. Install dependencies into a new conda environment

    conda install -n dask-cudf \
       -c rapidsai -c numba -c conda-forge -c defaults \
       pygdf dask distributed cudatoolkit
    
  2. Activate conda environment:

    source activate dask-cudf
    
  3. Clone dask_gdf repo:

    git clone https://github.com/rapidsai/dask-cudf
    
  4. Install from source:

    cd dask-cudf
    pip install .
    

Test

  1. Install pytest

    conda install pytest
    
  2. Run all tests:

    py.test dask_cudf
    
  3. Or, run individual tests:

     pytest dask_cudf/tests/test_file.py