RAPIDS: GPU Accelerated Data Science

Join us at the Keystone CoLAB to connect with Kansas City's data science community. Dr. Bradley Dice, Senior Software Engineer in GPU-Accelerated Data Analytics at NVIDIA will share about GPU Accelerated Data Science with RAPIDS.

The PyData ecosystem provides popular and powerful libraries like pandas, scikit-learn, and NetworkX for data analytics, machine learning, and graph analytics. The open-source RAPIDS libraries from NVIDIA bring accelerated computing to these familiar APIs in the PyData ecosystem, with GPU libraries like cuDF, cuML, and cuGraph. With the recent releases of cuDF pandas accelerator mode and NetworkX backend dispatching, it is now even possible to harness GPU acceleration and get significant performance improvements with no code changes at all. In this talk, Bradley will introduce the basics of using GPUs in data science and demonstrate how RAPIDS can turbocharge analytics workflows at any scale, from a single GPU laptop or workstation to an entire GPU cluster via dask or Spark.

Refreshments will be provided by Burns & McDonnell.