1 \select@language {french}
\r
3 \contentsline {figure}{\numberline {1.1}{\ignorespaces D\IeC {\'e}coupage d'une matrice tridimensionnelle entre deux clusters form\IeC {\'e}s de 18 processeurs chacun}}{8}{figure.1.1}
\r
4 \contentsline {figure}{\numberline {1.2}{\ignorespaces D\IeC {\'e}composition en domaines 1D, 2D et 3D}}{8}{figure.1.2}
\r
5 \contentsline {figure}{\numberline {1.3}{\ignorespaces Mod\IeC {\`e}le de communication synchrone}}{9}{figure.1.3}
\r
6 \contentsline {figure}{\numberline {1.4}{\ignorespaces Mod\IeC {\`e}le de communication asynchrone}}{9}{figure.1.4}
\r
7 \contentsline {figure}{\numberline {1.5}{\ignorespaces Algorithme it\IeC {\'e}ratif de Jacobi}}{12}{figure.1.5}
\r
8 \contentsline {figure}{\numberline {1.6}{\ignorespaces Algorithme it\IeC {\'e}ratif GMRES avec red\IeC {\'e}marrage}}{14}{figure.1.6}
\r
9 \contentsline {figure}{\numberline {1.7}{\ignorespaces Solveur Multisplitting utilisant la m\IeC {\'e}thode GMRES en local (version parall\IeC {\`e}le)}}{16}{figure.1.7}
\r
10 \contentsline {figure}{\numberline {1.8}{\ignorespaces Architecture d'une grille de calcul}}{17}{figure.1.8}
\r
11 \contentsline {figure}{\numberline {1.9}{\ignorespaces Grid'5000 : R\IeC {\'e}partition g\IeC {\'e}ographique}}{18}{figure.1.9}
\r
12 \contentsline {figure}{\numberline {1.10}{\ignorespaces Groupes et communicateur (a) - MPI - Op\IeC {\'e}rations collectives (b)}}{20}{figure.1.10}
\r
13 \contentsline {figure}{\numberline {1.11}{\ignorespaces SIMGRID : Les \IeC {\'e}l\IeC {\'e}ments de la plateforme de simulation}}{22}{figure.1.11}
\r
15 \contentsline {figure}{\numberline {2.1}{\ignorespaces Weak vs Strong scaling: Temps d'ex\IeC {\'e}cution et Speedup}}{27}{figure.2.1}
\r
16 \contentsline {figure}{\numberline {2.2}{\ignorespaces Architecture des CPU multicoeurs}}{31}{figure.2.2}
\r
17 \contentsline {figure}{\numberline {2.3}{\ignorespaces Mod\IeC {\`e}le MIMD M\IeC {\'e}moire Distribu\IeC {\'e}}}{32}{figure.2.3}
\r
18 \contentsline {figure}{\numberline {2.4}{\ignorespaces Mod\IeC {\`e}le MIMD M\IeC {\'e}moire partag\IeC {\'e}}}{33}{figure.2.4}
\r
19 \contentsline {figure}{\numberline {2.5}{\ignorespaces Mod\IeC {\`e}le MIMD hybride}}{34}{figure.2.5}
\r
20 \contentsline {figure}{\numberline {2.6}{\ignorespaces Evolution de la puissance de calcul mondiale}}{35}{figure.2.6}
\r
21 \contentsline {figure}{\numberline {2.7}{\ignorespaces M\IeC {\'e}moire MIMD: Architecture UMA}}{36}{figure.2.7}
\r
22 \contentsline {figure}{\numberline {2.8}{\ignorespaces M\IeC {\'e}moire MIMD: Architecture NUMA}}{36}{figure.2.8}
\r
23 \contentsline {figure}{\numberline {2.9}{\ignorespaces M\IeC {\'e}moire MIMD: Architecture COMA}}{37}{figure.2.9}
\r
24 \contentsline {figure}{\numberline {2.10}{\ignorespaces Classification des techniques d'analyse de la performance}}{39}{figure.2.10}
\r
