Administrator Guide

Terminology
19 RAPIDS Scaling on Dell EMC PowerEdge Servers
B Terminology
RAPIDS: Suite of software libraries, built on CUDA-X AI, that gives the freedom to execute end-to-end
data science and analytics pipelines entirely on GPUs
End to End workflow: Data science pipeline that includes the three phases of ETL (Extract, Transform,
Load), data conversion, and training
Dask: Open source freely available that provides advanced parallelism for analytics. It is developed in
coordination with other community projects like Numpy, Pandas, and Scikit-Learn
XGBoost: Open-source software library which provides a gradient boosting framework for C++, Java,
Python, R, and Julia. It works on Linux, Windows, and macOS
Docker mounted volume: An existing directory on the host that is “mounted” to be available inside the
container, useful for sharing files between the host and the container
Cluster: Group of computers communicating through fast interconnection
Node: Group of processors communicating through shared memory
Socket: Group of cores communicating through shared cache
Core: Group of functional units communicating through registers
Pipeline: Sequence of instructions sharing functional units
Threads: The smallest sequence of programmed instructions that can be managed independently by a
scheduler, which is typically a part of the operating system
Dask-scheduler: Coordinates and execute the task graphs on parallel hardware
Dask-worker: Computes tasks as directed by the schedules, stores and serves computed results to other
workers or clients
Dask-cuda: Allows deployment and management of Dask workers on CUDA-enabled systems
Diagnostic dashboard: Interactive dashboard containing several plots and tables with live information
about task runtimes, communication, statistical profiling, load balancing, memory use, and so on.
Bokeh: Interactive visualization library that targets modern web browsers for presentation