Changeset 4475

Timestamp:

Feb 7, 2015, 12:50:46 AM (4 years ago)

Author:

cameron

Message:

Discussion of performance results

File:

• docs/Working/icGrep/evaluation.tex (modified) (3 diffs)

docs/Working/icGrep/evaluation.tex

@@ -37 +37 @@
 most of the world's major language families as a test corpus.   For each program
 under test, we perform searches for each regular expression against each XML document.
-Searches were repeated n times.  Table \ref{tbl:property_test} shows the results.
-
-\begin{table}
-\input{table-prop.tex}
-\caption{Performance of Matching Property and Property Combinations}\label{tbl:property_test}
-\end{table}
+Results are presented in \Figure \ref{fig:property_test}.  Performance is reported
+in CPU cycles per byte on an Intel Core i7 machine.   The results were grouped
+by the percentage of matching lines found in the XML document, grouped in
+5\% increments.  ICgrep shows dramatically better performance, particularly
+when searching for rare items.
+As shown in the figure, pcre2grep and ugrep both show 
+increased performance (reduced CPU cycles per byte) with increasing percentage
+of matches found.  In essence, each match found allows these programs
+to skip the full processing of the rest of the line.   On the other
+hand, icGrep shows a slight drop-off in performance with the number
+of matches found.   This is primarily due to property classes that
+include large numbers of codepoints.   These classes require more
+bitstream equations for calculation and also have a greater probability
+of matching.   Nevertheless, the performance of icGrep in matching
+the defined property expressions is stable and well ahead of the competitors
+in all cases.
 
 \begin{figure}
+\begin{center}
 \pgfplotstableread[col sep = comma]{data/icgrep-scatter.csv}\icgrep
 \pgfplotstableread[col sep = comma]{data/ugrep541-scatter.csv}\ugrep
@@ -54 +65 @@
 x tick label style={ /pgf/number format/1000 sep=}, 
 % x buffer=sort,
-ylabel={Cycles Per Byte}, 
-xlabel={Match Percentage}, 
+ylabel={CPU Cycles Per Byte}, 
+xlabel={Percentage of Matching Lines}, 
 minor y tick num={1},
 xmax=100
@@ -68 +79 @@
 
 \end{tikzpicture}
-
+\end{center}
+\caption{Comparative Matching Performance}\label{fig:property_test}
 \end{figure}