deletions | additions       diff --git a/2-Parallelism.tex b/2-Parallelism.tex  index 2a1fb40..bff4bee 100644  --- a/2-Parallelism.tex  +++ b/2-Parallelism.tex 
...
\subsection{Parallel Batch Processing}  \label{sec:parallel}  Many \code{gadfly} use cases require Investigating the results of a simulation often requires  running an identical analysis on many snapshot in order to study how the system changes with time. These operations are often typically  completely independent, and thus amenable to parallelization so parallelization---so  long as the machine being used has sufficient memory. To aid in the parallelization of such analyses, \code{gadfly} includes utilities for performingsuch  parallel batch processing jobs by making use of python's \code{multiprocessing} module. One issue with doing operations such as this in parallel is that they are often IO-bound processes. IO-bound.  To mitigate the issues that arise when multiple processes attempt to read large amounts of data from disk simultaneously, \code{gadfly}'s parallelization utilities are designed to load the data needed for a given analysis serially in a separate thread from the main execution.   As the necessary data from each snapshot file is loaded, execution is passed to a pool of worker processes which can then perform the remainder of the analysis in parallel. An example of such a parallel batch processing script is included in the examples distributed with \code{gadfly}.