\@writefile{toc}{\author{Rapha\IeC {\"e}l Couturier}{}}
\@writefile{toc}{\author{Jacques Bahi}{}}
\@writefile{loa}{\addvspace {10\p@ }}
-\@writefile{toc}{\contentsline {chapter}{\numberline {10}Solving sparse linear systems with GMRES and CG methods on GPU clusters}{215}}
+\@writefile{toc}{\contentsline {chapter}{\numberline {11}Solving sparse linear systems with GMRES and CG methods on GPU clusters}{259}}
\@writefile{lof}{\addvspace {10\p@ }}
\@writefile{lot}{\addvspace {10\p@ }}
-\newlabel{ch12}{{10}{215}}
-\@writefile{toc}{\contentsline {section}{\numberline {10.1}Introduction}{215}}
-\newlabel{ch12:sec:01}{{10.1}{215}}
-\@writefile{toc}{\contentsline {section}{\numberline {10.2}Krylov iterative methods}{216}}
-\newlabel{ch12:sec:02}{{10.2}{216}}
-\newlabel{ch12:eq:01}{{10.1}{216}}
-\newlabel{ch12:eq:02}{{10.2}{216}}
-\newlabel{ch12:eq:03}{{10.3}{216}}
-\newlabel{ch12:eq:11}{{10.4}{217}}
-\@writefile{toc}{\contentsline {subsection}{\numberline {10.2.1}CG method}{217}}
-\newlabel{ch12:sec:02.01}{{10.2.1}{217}}
-\newlabel{ch12:eq:04}{{10.5}{217}}
-\newlabel{ch12:eq:05}{{10.6}{217}}
-\newlabel{ch12:eq:06}{{10.7}{217}}
-\newlabel{ch12:eq:07}{{10.8}{217}}
-\newlabel{ch12:eq:08}{{10.9}{217}}
-\newlabel{ch12:eq:09}{{10.10}{217}}
-\@writefile{loa}{\contentsline {algocf}{\numberline {9}{\ignorespaces Left-preconditioned CG method\relax }}{218}}
-\newlabel{ch12:alg:01}{{9}{218}}
-\newlabel{ch12:eq:10}{{10.11}{218}}
-\@writefile{toc}{\contentsline {subsection}{\numberline {10.2.2}GMRES method}{219}}
-\newlabel{ch12:sec:02.02}{{10.2.2}{219}}
-\newlabel{ch12:eq:12}{{10.12}{219}}
-\newlabel{ch12:eq:13}{{10.13}{219}}
-\newlabel{ch12:eq:14}{{10.14}{219}}
-\newlabel{ch12:eq:15}{{10.15}{219}}
-\newlabel{ch12:eq:16}{{10.16}{219}}
-\newlabel{ch12:eq:17}{{10.17}{219}}
-\newlabel{ch12:eq:18}{{10.18}{219}}
-\newlabel{ch12:eq:19}{{10.19}{219}}
-\@writefile{loa}{\contentsline {algocf}{\numberline {10}{\ignorespaces Left-preconditioned GMRES method with restarts\relax }}{220}}
-\newlabel{ch12:alg:02}{{10}{220}}
-\@writefile{toc}{\contentsline {section}{\numberline {10.3}Parallel implementation on a GPU cluster}{221}}
-\newlabel{ch12:sec:03}{{10.3}{221}}
-\@writefile{toc}{\contentsline {subsection}{\numberline {10.3.1}Data partitioning}{221}}
-\newlabel{ch12:sec:03.01}{{10.3.1}{221}}
-\@writefile{lof}{\contentsline {figure}{\numberline {10.1}{\ignorespaces A data partitioning of the sparse matrix $A$, the solution vector $x$ and the right-hand side $b$ into four portions.\relax }}{222}}
-\newlabel{ch12:fig:01}{{10.1}{222}}
-\@writefile{toc}{\contentsline {subsection}{\numberline {10.3.2}GPU computing}{222}}
-\newlabel{ch12:sec:03.02}{{10.3.2}{222}}
-\@writefile{toc}{\contentsline {subsection}{\numberline {10.3.3}Data communications}{223}}
-\newlabel{ch12:sec:03.03}{{10.3.3}{223}}
-\@writefile{lof}{\contentsline {figure}{\numberline {10.2}{\ignorespaces Data exchanges between \textit {Node 1} and its neighbors \textit {Node 0}, \textit {Node 2} and \textit {Node 3}.\relax }}{224}}
-\newlabel{ch12:fig:02}{{10.2}{224}}
-\@writefile{lof}{\contentsline {figure}{\numberline {10.3}{\ignorespaces Columns reordering of a sparse sub-matrix.\relax }}{225}}
-\newlabel{ch12:fig:03}{{10.3}{225}}
-\@writefile{toc}{\contentsline {section}{\numberline {10.4}Experimental results}{226}}
-\newlabel{ch12:sec:04}{{10.4}{226}}
-\@writefile{lof}{\contentsline {figure}{\numberline {10.4}{\ignorespaces General scheme of the GPU cluster of tests composed of six machines, each with two GPUs.\relax }}{226}}
-\newlabel{ch12:fig:04}{{10.4}{226}}
-\@writefile{lof}{\contentsline {figure}{\numberline {10.5}{\ignorespaces Sketches of sparse matrices chosen from the Davis collection.\relax }}{227}}
-\newlabel{ch12:fig:05}{{10.5}{227}}
-\@writefile{lot}{\contentsline {table}{\numberline {10.1}{\ignorespaces Main characteristics of sparse matrices chosen from the Davis collection.\relax }}{227}}
-\newlabel{ch12:tab:01}{{10.1}{227}}
-\@writefile{lot}{\contentsline {table}{\numberline {10.2}{\ignorespaces Performances of the parallel CG method on a cluster of 24 CPU cores vs. on a cluster of 12 GPUs.\relax }}{228}}
-\newlabel{ch12:tab:02}{{10.2}{228}}
-\@writefile{lot}{\contentsline {table}{\numberline {10.3}{\ignorespaces Performances of the parallel GMRES method on a cluster 24 CPU cores vs. on cluster of 12 GPUs.\relax }}{228}}
-\newlabel{ch12:tab:03}{{10.3}{228}}
-\newlabel{ch12:eq:20}{{10.20}{229}}
-\@writefile{lof}{\contentsline {figure}{\numberline {10.6}{\ignorespaces Parallel generation of a large sparse matrix by four computing nodes.\relax }}{230}}
-\newlabel{ch12:fig:06}{{10.6}{230}}
-\@writefile{lot}{\contentsline {table}{\numberline {10.4}{\ignorespaces Main characteristics of sparse banded matrices generated from those of the Davis collection.\relax }}{230}}
-\newlabel{ch12:tab:04}{{10.4}{230}}
-\@writefile{lot}{\contentsline {table}{\numberline {10.5}{\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 }}{231}}
-\newlabel{ch12:tab:05}{{10.5}{231}}
-\@writefile{toc}{\contentsline {section}{\numberline {10.5}Conclusion}{231}}
-\newlabel{ch12:sec:05}{{10.5}{231}}
-\@writefile{lot}{\contentsline {table}{\numberline {10.6}{\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 }}{232}}
-\newlabel{ch12:tab:06}{{10.6}{232}}
-\@writefile{toc}{\contentsline {section}{Bibliography}{232}}
+\newlabel{ch12}{{11}{259}}
+\@writefile{toc}{\contentsline {section}{\numberline {11.1}Introduction}{259}}
+\newlabel{ch12:sec:01}{{11.1}{259}}
+\@writefile{toc}{\contentsline {section}{\numberline {11.2}Krylov iterative methods}{260}}
+\newlabel{ch12:sec:02}{{11.2}{260}}
+\newlabel{ch12:eq:01}{{11.1}{260}}
+\newlabel{ch12:eq:02}{{11.2}{260}}
+\newlabel{ch12:eq:03}{{11.3}{260}}
+\newlabel{ch12:eq:11}{{11.4}{261}}
+\@writefile{toc}{\contentsline {subsection}{\numberline {11.2.1}CG method}{261}}
+\newlabel{ch12:sec:02.01}{{11.2.1}{261}}
+\newlabel{ch12:eq:04}{{11.5}{261}}
+\newlabel{ch12:eq:05}{{11.6}{261}}
+\newlabel{ch12:eq:06}{{11.7}{261}}
+\newlabel{ch12:eq:07}{{11.8}{261}}
+\newlabel{ch12:eq:08}{{11.9}{261}}
+\newlabel{ch12:eq:09}{{11.10}{261}}
+\@writefile{loa}{\contentsline {algocf}{\numberline {9}{\ignorespaces Left-preconditioned CG method\relax }}{262}}
+\newlabel{ch12:alg:01}{{9}{262}}
+\newlabel{ch12:eq:10}{{11.11}{262}}
+\@writefile{toc}{\contentsline {subsection}{\numberline {11.2.2}GMRES method}{263}}
+\newlabel{ch12:sec:02.02}{{11.2.2}{263}}
+\newlabel{ch12:eq:12}{{11.12}{263}}
+\newlabel{ch12:eq:13}{{11.13}{263}}
+\newlabel{ch12:eq:14}{{11.14}{263}}
+\newlabel{ch12:eq:15}{{11.15}{263}}
+\newlabel{ch12:eq:16}{{11.16}{263}}
+\newlabel{ch12:eq:17}{{11.17}{263}}
+\newlabel{ch12:eq:18}{{11.18}{263}}
+\newlabel{ch12:eq:19}{{11.19}{263}}
+\@writefile{loa}{\contentsline {algocf}{\numberline {10}{\ignorespaces Left-preconditioned GMRES method with restarts\relax }}{264}}
+\newlabel{ch12:alg:02}{{10}{264}}
+\@writefile{toc}{\contentsline {section}{\numberline {11.3}Parallel implementation on a GPU cluster}{265}}
+\newlabel{ch12:sec:03}{{11.3}{265}}
+\@writefile{toc}{\contentsline {subsection}{\numberline {11.3.1}Data partitioning}{265}}
+\newlabel{ch12:sec:03.01}{{11.3.1}{265}}
+\@writefile{lof}{\contentsline {figure}{\numberline {11.1}{\ignorespaces A data partitioning of the sparse matrix $A$, the solution vector $x$ and the right-hand side $b$ into four portions.\relax }}{266}}
+\newlabel{ch12:fig:01}{{11.1}{266}}
+\@writefile{toc}{\contentsline {subsection}{\numberline {11.3.2}GPU computing}{266}}
+\newlabel{ch12:sec:03.02}{{11.3.2}{266}}
+\@writefile{toc}{\contentsline {subsection}{\numberline {11.3.3}Data communications}{267}}
+\newlabel{ch12:sec:03.03}{{11.3.3}{267}}
+\@writefile{lof}{\contentsline {figure}{\numberline {11.2}{\ignorespaces Data exchanges between \textit {Node 1} and its neighbors \textit {Node 0}, \textit {Node 2} and \textit {Node 3}.\relax }}{268}}
+\newlabel{ch12:fig:02}{{11.2}{268}}
+\@writefile{lof}{\contentsline {figure}{\numberline {11.3}{\ignorespaces Columns reordering of a sparse sub-matrix.\relax }}{269}}
+\newlabel{ch12:fig:03}{{11.3}{269}}
+\@writefile{toc}{\contentsline {section}{\numberline {11.4}Experimental results}{270}}
+\newlabel{ch12:sec:04}{{11.4}{270}}
+\@writefile{lof}{\contentsline {figure}{\numberline {11.4}{\ignorespaces General scheme of the GPU cluster of tests composed of six machines, each with two GPUs.\relax }}{270}}
+\newlabel{ch12:fig:04}{{11.4}{270}}
+\@writefile{lof}{\contentsline {figure}{\numberline {11.5}{\ignorespaces Sketches of sparse matrices chosen from the Davis collection.\relax }}{271}}
+\newlabel{ch12:fig:05}{{11.5}{271}}
+\@writefile{lot}{\contentsline {table}{\numberline {11.1}{\ignorespaces Main characteristics of sparse matrices chosen from the Davis collection.\relax }}{271}}
+\newlabel{ch12:tab:01}{{11.1}{271}}
+\@writefile{lot}{\contentsline {table}{\numberline {11.2}{\ignorespaces Performances of the parallel CG method on a cluster of 24 CPU cores vs. on a cluster of 12 GPUs.\relax }}{272}}
+\newlabel{ch12:tab:02}{{11.2}{272}}
+\@writefile{lot}{\contentsline {table}{\numberline {11.3}{\ignorespaces Performances of the parallel GMRES method on a cluster 24 CPU cores vs. on cluster of 12 GPUs.\relax }}{272}}
+\newlabel{ch12:tab:03}{{11.3}{272}}
+\newlabel{ch12:eq:20}{{11.20}{273}}
+\@writefile{lof}{\contentsline {figure}{\numberline {11.6}{\ignorespaces Parallel generation of a large sparse matrix by four computing nodes.\relax }}{274}}
+\newlabel{ch12:fig:06}{{11.6}{274}}
+\@writefile{lot}{\contentsline {table}{\numberline {11.4}{\ignorespaces Main characteristics of sparse banded matrices generated from those of the Davis collection.\relax }}{274}}
+\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{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}{234}
+\setcounter{page}{278}
\setcounter{equation}{22}
\setcounter{enumi}{4}
\setcounter{enumii}{0}
\setcounter{footnote}{0}
\setcounter{mpfootnote}{0}
\setcounter{part}{5}
-\setcounter{chapter}{10}
+\setcounter{chapter}{11}
\setcounter{section}{5}
\setcounter{subsection}{0}
\setcounter{subsubsection}{0}
\setcounter{figure}{6}
\setcounter{table}{6}
\setcounter{numauthors}{0}
-\setcounter{parentequation}{8}
+\setcounter{parentequation}{46}
\setcounter{subfigure}{0}
\setcounter{lofdepth}{1}
\setcounter{subtable}{0}
