QA: Why do the RevoScaleR analysis functions typically run faster the second time?

The operating system caches recently-used data in RAM. On the second run or second iteration of algorithm used by an analysis function, access to the data is often much faster because some or all of the data is cached in RAM. 

The extent to which this is noticable depends upon the relative speeds of the hard drive, the number of CPUs, the amount of RAM available, and the the number of calculations performed in the CPU. 

Article ID: 3104272 - Last Review: Oct 29, 2015 - Revision: 1

Revolution Analytics