Skip to content

Latest commit

 

History

History
16 lines (11 loc) · 782 Bytes

README.md

File metadata and controls

16 lines (11 loc) · 782 Bytes

Benchmarking

This directory contains python scripts for benchmarking the supported algorithms.

Local

This script can be used to run them locally.

Databricks

They can also be run on the Databricks AWS hosted Spark service. See these instructions and accompanying scripts for running a set of high compute workloads on comparable CPU and GPU clusters. The graph below shows the resulting Spark ML CPU and Spark Rapids ML GPU average running times.

Databricks AWS benchmarking results

Other CSPs

Click on the below for instructions on running the benchmarking scripts in the respective CSP Spark environments