The request names cuPyNumeric, whose parallel data loading is served by the NVIDIA product's action 'cupynumeric-parallel-data-load', which loads sharded on-disk datasets into a distributed cuPyNumeric ndarray 111. Related actions for cuPyNumeric HDF5 I/O and installation are also indexed under the same record 111. Alternatives that handle parallel data loading in Python are Dask 3 and Polars 6, though their sources do not name cuPyNumeric.

1. 1

   NVIDIA cuPyNumericnvidia.comverified and signed

   Use the action cupynumeric-parallel-data-load to load a sharded on-disk dataset (sharded .npy, Parquet/Arrow, raw binary, sharded HDF5, custom layouts) into a distributed cuPyNumeric ndarray 11.

   WebsiteDocsPricingGet started

   +84 community tools built for it

   for agents: cupynumeric-parallel-data-load · cupynumeric-hdf5 · cupynumeric-install

2. 2

   Daskdask.orgverified and signed

   A flexible open-source Python library for parallel computing that scales Python code across machines, with DataFrames built on pandas 313.

   WebsiteDocsGet started

   +2 community tools built for it

   for agents: parallel · numpy · pydata

3. 3

   Polarspola.rsverified and signed

   A high-performance DataFrame library for Python and Rust with an out-of-core engine for tabular data at scale 616.

   WebsiteDocs

   +1 community tools built for it

   for agents: dataframe · dataframe · out-of-core