From: couturie Date: Wed, 27 Mar 2013 21:01:44 +0000 (+0100) Subject: ch19 X-Git-Url: https://bilbo.iut-bm.univ-fcomte.fr/and/gitweb/book_gpu.git/commitdiff_plain/ad42c5f71154310efeb0bb918ec90ee38f421641?ds=sidebyside;hp=-c ch19 --- ad42c5f71154310efeb0bb918ec90ee38f421641 diff --git a/BookGPU/BookGPU.tex b/BookGPU/BookGPU.tex index de8e359..c038681 100755 --- a/BookGPU/BookGPU.tex +++ b/BookGPU/BookGPU.tex @@ -34,6 +34,7 @@ \usepackage{stmaryrd} \usepackage{moreverb} \usepackage{commath} +\usepackage{numprint} %% \usepackage{listings} %% \usepackage{subfigure} @@ -186,6 +187,7 @@ \include{Chapters/chapter16/ch16} \part{Other} \include{Chapters/chapter18/ch18} +\include{Chapters/chapter19/ch19} \bibliographystyle{hep} %%%\bibliography{biblio} diff --git a/BookGPU/Chapters/chapter12/ch12.aux b/BookGPU/Chapters/chapter12/ch12.aux index 82783b4..2252156 100644 --- a/BookGPU/Chapters/chapter12/ch12.aux +++ b/BookGPU/Chapters/chapter12/ch12.aux @@ -71,10 +71,10 @@ \newlabel{ch12:tab:04}{{11.4}{274}} \@writefile{lot}{\contentsline {table}{\numberline {11.5}{\ignorespaces Performances of the parallel CG method for solving linear systems associated to sparse banded matrices on a cluster of 24 CPU cores vs. on a cluster of 12 GPUs.\relax }}{275}} \newlabel{ch12:tab:05}{{11.5}{275}} -\@writefile{lot}{\contentsline {table}{\numberline {11.6}{\ignorespaces Performances of the parallel GMRES method for solving linear systems associated to sparse banded matrices on a cluster of 24 CPU cores vs. on a cluster of 12 GPUs.\relax }}{275}} -\newlabel{ch12:tab:06}{{11.6}{275}} \@writefile{toc}{\contentsline {section}{\numberline {11.5}Conclusion}{275}} \newlabel{ch12:sec:05}{{11.5}{275}} +\@writefile{lot}{\contentsline {table}{\numberline {11.6}{\ignorespaces Performances of the parallel GMRES method for solving linear systems associated to sparse banded matrices on a cluster of 24 CPU cores vs. on a cluster of 12 GPUs.\relax }}{276}} +\newlabel{ch12:tab:06}{{11.6}{276}} \@writefile{toc}{\contentsline {section}{Bibliography}{276}} \@setckpt{Chapters/chapter12/ch12}{ \setcounter{page}{278} @@ -106,6 +106,13 @@ \setcounter{algocfline}{10} \setcounter{algocfproc}{10} \setcounter{algocf}{10} +\setcounter{nprt@mantissa@digitsbefore}{0} +\setcounter{nprt@mantissa@digitsafter}{0} +\setcounter{nprt@exponent@digitsbefore}{0} +\setcounter{nprt@exponent@digitsafter}{0} +\setcounter{nprt@digitsfirstblock}{0} +\setcounter{nprt@blockcnt}{0} +\setcounter{nprt@cntprint}{0} \setcounter{proposition}{1} \setcounter{theorem}{0} \setcounter{exercise}{0} diff --git a/BookGPU/Chapters/chapter16/ch16.aux b/BookGPU/Chapters/chapter16/ch16.aux index 75c0afb..a0d90ea 100644 --- a/BookGPU/Chapters/chapter16/ch16.aux +++ b/BookGPU/Chapters/chapter16/ch16.aux @@ -98,6 +98,13 @@ \setcounter{algocfline}{15} \setcounter{algocfproc}{15} \setcounter{algocf}{15} +\setcounter{nprt@mantissa@digitsbefore}{0} +\setcounter{nprt@mantissa@digitsafter}{0} +\setcounter{nprt@exponent@digitsbefore}{0} +\setcounter{nprt@exponent@digitsafter}{0} +\setcounter{nprt@digitsfirstblock}{0} +\setcounter{nprt@blockcnt}{0} +\setcounter{nprt@cntprint}{0} \setcounter{proposition}{1} \setcounter{theorem}{0} \setcounter{exercise}{0} diff --git a/BookGPU/Chapters/chapter18/ch18.aux b/BookGPU/Chapters/chapter18/ch18.aux index 618a140..b69311d 100644 --- a/BookGPU/Chapters/chapter18/ch18.aux +++ b/BookGPU/Chapters/chapter18/ch18.aux @@ -69,6 +69,13 @@ \setcounter{algocfline}{17} \setcounter{algocfproc}{17} \setcounter{algocf}{17} +\setcounter{nprt@mantissa@digitsbefore}{0} +\setcounter{nprt@mantissa@digitsafter}{0} +\setcounter{nprt@exponent@digitsbefore}{0} +\setcounter{nprt@exponent@digitsafter}{0} +\setcounter{nprt@digitsfirstblock}{0} +\setcounter{nprt@blockcnt}{0} +\setcounter{nprt@cntprint}{0} \setcounter{proposition}{1} \setcounter{theorem}{0} \setcounter{exercise}{0} diff --git a/BookGPU/Chapters/chapter3/ch3.aux b/BookGPU/Chapters/chapter3/ch3.aux index f86e67e..298a450 100644 --- a/BookGPU/Chapters/chapter3/ch3.aux +++ b/BookGPU/Chapters/chapter3/ch3.aux @@ -62,15 +62,15 @@ \newlabel{lst:kernelMedian3RegTri9}{{4.2}{38}} \@writefile{lol}{\contentsline {lstlisting}{\numberline {4.2}3$\times $3 median filter kernel using one register per neighborhood pixel and bubble sort}{38}} \@writefile{toc}{\contentsline {subsection}{\numberline {4.4.2}Further optimization}{38}} -\@writefile{toc}{\contentsline {subsubsection}{\numberline {4.4.2.1}Reducing register count }{38}} \@writefile{lof}{\contentsline {figure}{\numberline {4.4}{\ignorespaces Comparison of pixel throughputs on GPU C2070 and CPU for generic median, 3$\times $3 median register-only and \textit {libJacket}.\relax }}{39}} \newlabel{fig:compMedians1}{{4.4}{39}} \@writefile{lof}{\contentsline {figure}{\numberline {4.5}{\ignorespaces Forgetful selection with the minimal element register count. Illustration for 3$\times $3 pixel window represented in a row and supposed sorted.\relax }}{39}} \newlabel{fig:forgetful_selection}{{4.5}{39}} -\@writefile{lof}{\contentsline {figure}{\numberline {4.6}{\ignorespaces Determination of the Median value by the forgetful selection process, applied to a $3\times 3$ neighborhood window.\relax }}{40}} -\newlabel{fig:forgetful3}{{4.6}{40}} -\newlabel{lst:medianForget1pix3}{{4.3}{41}} -\@writefile{lol}{\contentsline {lstlisting}{\numberline {4.3}3$\times $3 median filter kernel using the minimum register count of 6 to find the median value by forgetful selection method. The optimal thread block size is 128 on GTX280 and 256 on C2070.}{41}} +\@writefile{toc}{\contentsline {subsubsection}{\numberline {4.4.2.1}Reducing register count }{39}} +\newlabel{lst:medianForget1pix3}{{4.3}{40}} +\@writefile{lol}{\contentsline {lstlisting}{\numberline {4.3}3$\times $3 median filter kernel using the minimum register count of 6 to find the median value by forgetful selection method. The optimal thread block size is 128 on GTX280 and 256 on C2070.}{40}} +\@writefile{lof}{\contentsline {figure}{\numberline {4.6}{\ignorespaces Determination of the Median value by the forgetful selection process, applied to a $3\times 3$ neighborhood window.\relax }}{41}} +\newlabel{fig:forgetful3}{{4.6}{41}} \@writefile{lof}{\contentsline {figure}{\numberline {4.7}{\ignorespaces Illustration of how window overlapping is used to combine 2 pixel selections in a 3$\times $3 median kernel.\relax }}{42}} \newlabel{fig:median3_overlap}{{4.7}{42}} \@writefile{toc}{\contentsline {subsubsection}{\numberline {4.4.2.2}More data output per thread}{42}} @@ -139,6 +139,13 @@ \setcounter{algocfline}{2} \setcounter{algocfproc}{2} \setcounter{algocf}{2} +\setcounter{nprt@mantissa@digitsbefore}{0} +\setcounter{nprt@mantissa@digitsafter}{0} +\setcounter{nprt@exponent@digitsbefore}{0} +\setcounter{nprt@exponent@digitsafter}{0} +\setcounter{nprt@digitsfirstblock}{0} +\setcounter{nprt@blockcnt}{0} +\setcounter{nprt@cntprint}{0} \setcounter{proposition}{0} \setcounter{theorem}{0} \setcounter{exercise}{0} diff --git a/BookGPU/Chapters/chapter6/ch6.aux b/BookGPU/Chapters/chapter6/ch6.aux index 1527dfc..bfe00ce 100644 --- a/BookGPU/Chapters/chapter6/ch6.aux +++ b/BookGPU/Chapters/chapter6/ch6.aux @@ -135,6 +135,13 @@ \setcounter{algocfline}{4} \setcounter{algocfproc}{4} \setcounter{algocf}{4} +\setcounter{nprt@mantissa@digitsbefore}{0} +\setcounter{nprt@mantissa@digitsafter}{0} +\setcounter{nprt@exponent@digitsbefore}{0} +\setcounter{nprt@exponent@digitsafter}{0} +\setcounter{nprt@digitsfirstblock}{0} +\setcounter{nprt@blockcnt}{0} +\setcounter{nprt@cntprint}{0} \setcounter{proposition}{0} \setcounter{theorem}{0} \setcounter{exercise}{0} diff --git a/BookGPU/Makefile b/BookGPU/Makefile index cdc3c59..9956a25 100644 --- a/BookGPU/Makefile +++ b/BookGPU/Makefile @@ -20,6 +20,7 @@ all: bibtex bu13 bibtex bu14 bibtex bu15 + bibtex bu16 makeindex ${BOOK}.idx pdflatex ${BOOK}