Merge pull request #239 from ldbc/cross-validation-minimum-number-of-operations

Specify minimum number of operations for cross-validation during auditing

Author: szarnyasg

auditing.tex

@@ -360,7 +360,8 @@ \subsubsection{Scale Factors}
 The size of a data set is characterized by scale factors: SF10, SF30, SF100 \etc (see \autoref{sec:scale-factors}).
 All data sets contain data for three years of social network activity.
 
-The \emph{validation run} shall be performed on the SF10 data set (see \autoref{sec:int-validation-data-set}). Note that the auditor may perform additional validation runs of the benchmark implementation using smaller data sets (\eg SF1) and issue queries.\footnote{%
+The \emph{validation run} shall be performed on the SF10 data set (see \autoref{sec:int-validation-data-set}) and use at least \numprint{100000} operations.
+Note that the auditor may perform additional validation runs of the benchmark implementation using smaller data sets (\eg SF1) and issue queries.\footnote{%
 An example test could be to issue complex reads with parameters such as \texttt{personId} and \texttt{messageId} selected from the \textsf{Person}/\textsf{Message} entities inserted from the update streams and cross-validate these against other systems. (The substitution parameters are taken from the initial snapshot of the graph so these nodes are not targeted by the regular workload executed by the driver.)%
 }