2 \@writefile{toc}{\author{}{}}
3 \@writefile{loa}{\addvspace {10\p@ }}
4 \@writefile{toc}{\contentsline {chapter}{\numberline {11}Solving sparse linear systems with GMRES and CG methods on GPU clusters}{249}}
5 \@writefile{lof}{\addvspace {10\p@ }}
6 \@writefile{lot}{\addvspace {10\p@ }}
7 \@writefile{toc}{\contentsline {section}{\numberline {11.1}Introduction}{249}}
8 \newlabel{sec:01}{{11.1}{249}}
9 \@writefile{toc}{\contentsline {section}{\numberline {11.2}Krylov iterative methods}{250}}
10 \newlabel{sec:02}{{11.2}{250}}
11 \newlabel{eq:01}{{11.1}{250}}
12 \newlabel{eq:02}{{11.2}{250}}
13 \newlabel{eq:03}{{11.3}{250}}
14 \newlabel{eq:11}{{11.4}{251}}
15 \@writefile{toc}{\contentsline {subsection}{\numberline {11.2.1}CG method}{251}}
16 \newlabel{sec:02.01}{{11.2.1}{251}}
17 \newlabel{eq:04}{{11.5}{251}}
18 \newlabel{eq:05}{{11.6}{251}}
19 \newlabel{eq:06}{{11.7}{251}}
20 \newlabel{eq:07}{{11.8}{251}}
21 \newlabel{eq:08}{{11.9}{251}}
22 \newlabel{eq:09}{{11.10}{251}}
23 \@writefile{loa}{\contentsline {algocf}{\numberline {9}{\ignorespaces Left-preconditioned CG method\relax }}{252}}
24 \newlabel{alg:01}{{9}{252}}
25 \newlabel{eq:10}{{11.11}{252}}
26 \@writefile{toc}{\contentsline {subsection}{\numberline {11.2.2}GMRES method}{253}}
27 \newlabel{sec:02.02}{{11.2.2}{253}}
28 \newlabel{eq:12}{{11.12}{253}}
29 \newlabel{eq:13}{{11.13}{253}}
30 \newlabel{eq:14}{{11.14}{253}}
31 \newlabel{eq:15}{{11.15}{253}}
32 \newlabel{eq:16}{{11.16}{253}}
33 \newlabel{eq:17}{{11.17}{253}}
34 \newlabel{eq:18}{{11.18}{253}}
35 \newlabel{eq:19}{{11.19}{253}}
36 \@writefile{loa}{\contentsline {algocf}{\numberline {10}{\ignorespaces Left-preconditioned GMRES method with restarts\relax }}{254}}
37 \newlabel{alg:02}{{10}{254}}
38 \@writefile{toc}{\contentsline {section}{\numberline {11.3}Parallel implementation on a GPU cluster}{255}}
39 \newlabel{sec:03}{{11.3}{255}}
40 \@writefile{toc}{\contentsline {subsection}{\numberline {11.3.1}Data partitioning}{255}}
41 \newlabel{sec:03.01}{{11.3.1}{255}}
42 \@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 }}{256}}
43 \newlabel{fig:01}{{11.1}{256}}
44 \@writefile{toc}{\contentsline {subsection}{\numberline {11.3.2}GPU computing}{256}}
45 \newlabel{sec:03.02}{{11.3.2}{256}}
46 \@writefile{toc}{\contentsline {subsection}{\numberline {11.3.3}Data communications}{257}}
47 \newlabel{sec:03.03}{{11.3.3}{257}}
48 \@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 }}{258}}
49 \newlabel{fig:02}{{11.2}{258}}
50 \@writefile{lof}{\contentsline {figure}{\numberline {11.3}{\ignorespaces Columns reordering of a sparse sub-matrix.\relax }}{259}}
51 \newlabel{fig:03}{{11.3}{259}}
52 \@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 }}{260}}
53 \newlabel{fig:04}{{11.4}{260}}
54 \@writefile{toc}{\contentsline {section}{\numberline {11.4}Experimental results}{260}}
55 \newlabel{sec:04}{{11.4}{260}}
56 \@writefile{lof}{\contentsline {figure}{\numberline {11.5}{\ignorespaces Sketches of sparse matrices chosen from the Davis's collection.\relax }}{261}}
57 \newlabel{fig:05}{{11.5}{261}}
58 \@writefile{lot}{\contentsline {table}{\numberline {11.1}{\ignorespaces Main characteristics of sparse matrices chosen from the Davis's collection.\relax }}{262}}
59 \newlabel{tab:01}{{11.1}{262}}
60 \@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 }}{262}}
61 \newlabel{tab:02}{{11.2}{262}}
62 \@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 }}{263}}
63 \newlabel{tab:03}{{11.3}{263}}
64 \newlabel{eq:20}{{11.20}{263}}
65 \@writefile{lof}{\contentsline {figure}{\numberline {11.6}{\ignorespaces Parallel generation of a large sparse matrix by four computing nodes.\relax }}{264}}
66 \newlabel{fig:06}{{11.6}{264}}
67 \@writefile{lot}{\contentsline {table}{\numberline {11.4}{\ignorespaces Main characteristics of sparse banded matrices generated from those of the Davis's collection.\relax }}{265}}
68 \newlabel{tab:04}{{11.4}{265}}
69 \@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 }}{265}}
70 \newlabel{tab:05}{{11.5}{265}}
71 \@writefile{toc}{\contentsline {section}{\numberline {11.5}Hypergraph partitioning}{265}}
72 \newlabel{sec:05}{{11.5}{265}}
73 \@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 }}{266}}
74 \newlabel{tab:06}{{11.6}{266}}
75 \@writefile{lot}{\contentsline {table}{\numberline {11.7}{\ignorespaces Main characteristics of sparse five-bands matrices generated from those of the Davis's collection.\relax }}{266}}
76 \newlabel{tab:07}{{11.7}{266}}
77 \@writefile{lof}{\contentsline {figure}{\numberline {11.7}{\ignorespaces Parallel generation of a large sparse five-bands matrix by four computing nodes.\relax }}{267}}
78 \newlabel{fig:07}{{11.7}{267}}
79 \@writefile{lot}{\contentsline {table}{\numberline {11.8}{\ignorespaces Performances of parallel CG solver for solving linear systems associated to sparse five-bands matrices on a cluster of 24 CPU cores vs. on a cluster of 12 GPUs\relax }}{267}}
80 \newlabel{tab:08}{{11.8}{267}}
81 \@writefile{lot}{\contentsline {table}{\numberline {11.9}{\ignorespaces Performances of parallel GMRES solver for solving linear systems associated to sparse five-bands matrices on a cluster of 24 CPU cores vs. on a cluster of 12 GPUs\relax }}{268}}
82 \newlabel{tab:09}{{11.9}{268}}
83 \@writefile{lof}{\contentsline {figure}{\numberline {11.8}{\ignorespaces An example of the hypergraph partitioning of a sparse matrix decomposed between three computing nodes.\relax }}{269}}
84 \newlabel{fig:08}{{11.8}{269}}
85 \@writefile{lot}{\contentsline {table}{\numberline {11.10}{\ignorespaces Performances of the parallel CG solver using hypergraph partitioning for solving linear systems associated to sparse five-bands matrices on a cluster of 24 CPU cores vs. on a cluster of 12 GPU.\relax }}{270}}
86 \newlabel{tab:10}{{11.10}{270}}
87 \@writefile{lot}{\contentsline {table}{\numberline {11.11}{\ignorespaces Performances of the parallel GMRES solver using hypergraph partitioning for solving linear systems associated to sparse five-bands matrices on a cluster of 24 CPU cores vs. on a cluster of 12 GPU.\relax }}{271}}
88 \newlabel{tab:11}{{11.11}{271}}
89 \@writefile{lot}{\contentsline {table}{\numberline {11.12}{\ignorespaces The total communication volume between 12 GPU computing nodes without and with the hypergraph partitioning method.\relax }}{272}}
90 \newlabel{tab:12}{{11.12}{272}}
91 \@writefile{lof}{\contentsline {figure}{\numberline {11.9}{\ignorespaces Weak-scaling of the parallel CG and GMRES solvers on a GPU cluster for solving large sparse linear systems.\relax }}{273}}
92 \newlabel{fig:09}{{11.9}{273}}
93 \@writefile{toc}{\contentsline {section}{\numberline {11.6}Conclusion}{273}}
94 \newlabel{sec:06}{{11.6}{273}}
95 \@writefile{toc}{\contentsline {section}{Bibliography}{274}}
96 \@setckpt{Chapters/chapter12/ch12}{
97 \setcounter{page}{276}
98 \setcounter{equation}{25}
100 \setcounter{enumii}{0}
101 \setcounter{enumiii}{0}
102 \setcounter{enumiv}{15}
103 \setcounter{footnote}{0}
104 \setcounter{mpfootnote}{0}
106 \setcounter{chapter}{11}
107 \setcounter{section}{6}
108 \setcounter{subsection}{0}
109 \setcounter{subsubsection}{0}
110 \setcounter{paragraph}{0}
111 \setcounter{subparagraph}{0}
112 \setcounter{figure}{9}
113 \setcounter{table}{12}
114 \setcounter{numauthors}{0}
115 \setcounter{parentequation}{46}
116 \setcounter{subfigure}{0}
117 \setcounter{lofdepth}{1}
118 \setcounter{subtable}{0}
119 \setcounter{lotdepth}{1}
120 \setcounter{lstnumber}{50}
121 \setcounter{ContinuedFloat}{0}
122 \setcounter{AlgoLine}{29}
123 \setcounter{algocfline}{10}
124 \setcounter{algocfproc}{10}
125 \setcounter{algocf}{10}
126 \setcounter{proposition}{1}
127 \setcounter{theorem}{0}
128 \setcounter{exercise}{0}
129 \setcounter{example}{0}
130 \setcounter{definition}{0}
131 \setcounter{proof}{1}
132 \setcounter{lstlisting}{0}
