diogo/yap-6.3
Archived
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Revised 7.1 and 7.2

Browse Source 
git-svn-id: https://yap.svn.sf.net/svnroot/yap/trunk@1832 b08c6af1-5177-4d33-ba66-4b1c6b8b522a
This commit is contained in:
kostis 2007-03-11 20:57:22 +00:00
parent 9ed8306415
commit 74da8a99fc
1 changed files with 24 additions and 23 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

47
docs/index/iclp07.tex
View File

@ -981,7 +981,7 @@ convenient abbreviation.
\section{Performance Evaluation} \label{sec:perf} \section{Performance Evaluation} \label{sec:perf}
%================================================ %================================================
We evaluate \JITI on a set of benchmarks and on applications. We evaluate \JITI on a set of benchmarks and logic programming applications.
Throughout, we compare performance of JITI with first argument Throughout, we compare performance of JITI with first argument
indexing. For the benchmarks of Sect.~\ref{sec:perf:ineffective} indexing. For the benchmarks of Sect.~\ref{sec:perf:ineffective}
and~\ref{sec:perf:effective} which involve both systems, we used a and~\ref{sec:perf:effective} which involve both systems, we used a
@ -1001,17 +1001,17 @@ have no effect other than an overhead due to runtime index
construction. We therefore wanted to measure this overhead. construction. We therefore wanted to measure this overhead.
% %
As both systems support tabling, we decided to use tabling benchmarks As both systems support tabling, we decided to use tabling benchmarks
because they are relatively small and easy to understand, and because because they are small and easy to understand, and because they are a
they are a worst case for JITI in the following sense: tabling avoids worst case for JITI in the following sense: tabling avoids generating
generating repetitive queries and the benchmarks operate over EDB repetitive queries and the benchmarks operate over EDB predicates of
predicates of size approximately equal the size of the program. size approximately equal the size of the program. We used \compress, a
We used \compress, a tabled program that solves a puzzle from an ICLP tabled program that solves a puzzle from an ICLP Prolog programming
Prolog programming competition. The other benchmarks are different competition. The other benchmarks are different variants of tabled
variants of tabled left, right and doubly recursive transitive closure left, right and doubly recursive transitive closure over an EDB
over an EDB predicate forming a chain of size shown in predicate forming a chain of size shown in Table~\ref{tab:ineffective}
Table~\ref{tab:ineffective} in parentheses. For each variant of in parentheses. For each variant of transitive closure, we issue two
transitive closure, we issue two queries: one with mode queries: one with mode \code{(in,out)} and one with mode
\code{(in,out)} and one with mode \code{(out,out)}. \code{(out,out)}.
% %
For YAP, indices on the first argument are built on all benchmarks For YAP, indices on the first argument are built on all benchmarks
under JITI.\TODO{Vitor please verify this sentence} under JITI.\TODO{Vitor please verify this sentence}
@ -1055,9 +1055,9 @@ columns separately.
\cline{2-7} \cline{2-7}
Benchmark & 1st & JITI &{\bf ratio}& 1st & JITI &{\bf ratio}\\ Benchmark & 1st & JITI &{\bf ratio}& 1st & JITI &{\bf ratio}\\
\hline \hline
\sgCyl & 2864 & 24 &$119\times$& 2390 & 28 &$85\times$\\ \sgCyl & 2,864 & 24 &$119\times$& 2,390 & 28 &$85\times$\\
\muta & 30,057 &16,782 &$179\%$ &26,314 &21,574 &$122\%$ \\ \muta & 30,057 &16,782 &$179\%$ &26,314 &21,574 &$122\%$ \\
\pta & 5131 & 188 & $27\times$& 4442 & 279 &$16\times$\\ \pta & 5,131 & 188 & $27\times$& 4,442 & 279 &$16\times$\\
\tea &1,478,813 &54,616 & $27\times$& --- & --- & --- \\ \tea &1,478,813 &54,616 & $27\times$& --- & --- & --- \\
\hline \hline
\end{tabular} \end{tabular}
@ -1068,9 +1068,9 @@ columns separately.
\subsection{Performance of \JITI when effective} \label{sec:perf:effective} \subsection{Performance of \JITI when effective} \label{sec:perf:effective}
%-------------------------------------------------------------------------- %--------------------------------------------------------------------------
On the other hand, when \JITI is effective, it can significantly On the other hand, when \JITI is effective, it can significantly
improve time performance. We use the following programs:\TODO{If time improve time performance. We use the following programs and
permits, we should also add FSA benchmarks (\bench{k963}, \bench{dg5} applications:\TODO{If time permits, we should also add FSA benchmarks
and \bench{tl3})} (\bench{k963}, \bench{dg5} and \bench{tl3})}
\begin{description} \begin{description}
\item[\sgCyl] The same generation DB benchmark on a $24 \times 24 \item[\sgCyl] The same generation DB benchmark on a $24 \times 24
\times 2$ cylinder. We issue the open query. \times 2$ cylinder. We issue the open query.
@ -1090,9 +1090,9 @@ and \bench{tl3})}
As can be seen in Table~\ref{tab:effective}, \JITI significantly As can be seen in Table~\ref{tab:effective}, \JITI significantly
improves the performance of these applications. In \muta, which spends improves the performance of these applications. In \muta, which spends
most of its time in recursive predicates, the speed up is~$79\%$ in most of its time in recursive predicates, the speed up is only $79\%$
YAP and~$22\%$ in XXX. The remaining benchmarks execute several times in YAP and~$22\%$ in XXX. The remaining benchmarks execute several
(from~$16$ up to~$119$) faster. It is important to realize that times (from~$16$ up to~$119$) faster. It is important to realize that
\emph{these speedups are obtained automatically}, i.e., without any \emph{these speedups are obtained automatically}, i.e., without any
programmer intervention or by using any compiler directives, in all programmer intervention or by using any compiler directives, in all
these applications. these applications.
@ -1103,8 +1103,8 @@ With the open call to \texttt{same\_generation/2}, most work in this
benchmark consists of calling \texttt{cyl/2} facts in three different benchmark consists of calling \texttt{cyl/2} facts in three different
modes: with both arguments unbound, with the first argument bound, or modes: with both arguments unbound, with the first argument bound, or
with only the second argument bound. Demand-driven indexing improves with only the second argument bound. Demand-driven indexing improves
performance in the last case only, but this makes a big difference in performance in the last case only, but this improvement makes a big
this benchmark. difference in this benchmark.
\begin{small} \begin{small}
\begin{verbatim} \begin{verbatim}
@ -1121,7 +1121,8 @@ this benchmark.
\subsection{Performance of \JITI on ILP applications} \label{sec:perf:ILP} \subsection{Performance of \JITI on ILP applications} \label{sec:perf:ILP}
%------------------------------------------------------------------------- %-------------------------------------------------------------------------
The need for \JITI was originally motivated by ILP applications. The need for \JITI was originally noticed in inductive logic
programming applications.
Table~\ref{tab:ilp:time} shows JITI performance on some learning tasks Table~\ref{tab:ilp:time} shows JITI performance on some learning tasks
using the ALEPH system~\cite{ALEPH}. The dataset \Krki tries to using the ALEPH system~\cite{ALEPH}. The dataset \Krki tries to
learn rules from a small database of chess end-games; learn rules from a small database of chess end-games;