文件列表:

AMG\V1.2\AccumulateMatrix.m (573, 2006-07-27)
AMG\V1.2\AMGLab.m (40, 2006-07-27)
AMG\V1.2\AMGLab_GUI.fig (3546, 2006-07-27)
AMG\V1.2\AMGLab_GUI.m (10239, 2006-07-27)
AMG\V1.2\AMGmCoarsen.m (17706, 2006-07-27)
AMG\V1.2\amg_cycle.m (2226, 2006-07-27)
AMG\V1.2\amg_defaults.m (542, 2006-07-27)
AMG\V1.2\amg_example1.m (2255, 2006-07-27)
AMG\V1.2\amg_globals.m (3861, 2006-07-27)
AMG\V1.2\amg_setup.m (768, 2006-07-27)
AMG\V1.2\amg_solve.m (3101, 2006-07-27)
AMG\V1.2\amg_usersetFD.m (579, 2006-07-27)
AMG\V1.2\amg_usersetFEM.m (1810, 2006-07-27)
AMG\V1.2\anisox_81.txt (79214, 2006-07-27)
AMG\V1.2\coarse_solve.m (348, 2006-07-27)
AMG\V1.2\CVS\Entries (1591, 2006-07-27)
AMG\V1.2\CVS\Repository (35, 2006-07-27)
AMG\V1.2\CVS\Root (46, 2006-07-27)
AMG\V1.2\e32_isb.txt (2077, 2006-07-27)
AMG\V1.2\gauss_elim.m (399, 2006-07-27)
AMG\V1.2\GetEdgeElements_Schur.m (3057, 2006-07-27)
AMG\V1.2\get_settings.m (3866, 2006-07-27)
AMG\V1.2\interpolate.m (195, 2006-07-27)
AMG\V1.2\jacobi_smoother.m (888, 2006-07-27)
AMG\V1.2\NodeNeighbors.m (606, 2006-07-27)
AMG\V1.2\NodeNeighborsEdges.m (765, 2006-07-27)
AMG\V1.2\residual.m (312, 2006-07-27)
AMG\V1.2\restrict.m (276, 2006-07-27)
AMG\V1.2\RugeStuebenCoarsen.m (5283, 2006-07-27)
AMG\V1.2\settingsGUI.fig (5688, 2006-07-27)
AMG\V1.2\settingsGUI.m (20190, 2006-07-27)
AMG\V1.2\smooth.m (377, 2006-07-27)
AMG\V1.2\SmoothAggregateCoarsen.m (7289, 2006-07-27)
AMG\V1.2\SOR_smoother.m (903, 2006-07-27)
AMG\V1.2\strong.m (1262, 2006-07-27)
AMG\V1.2\strongT.m (798, 2006-07-27)
AMG\V1.2\CVS (0, 2008-09-21)
AMG\V1.2 (0, 2008-09-21)
AMG (0, 2008-09-21)
... ...

Release V1.2 07 26, 2006 Features: * New to this release July 10, 2006 (appx 4:30pm EST) We are working on the plotting of the residual reduction. The way I was keeping track of the residual history was before and after correction at each level. Since this is not full multigrid, the courser levels are solving the residual correction of the finer levels, NOT the original problem. Thus, the way I was plotting residual reduction before did not make sense... More changes we thought of today... Add more cycle types other than the simple V cycle Add ability for AMG cycle to automatically determine how many cycles/levels to execute ------------------------------------------------------------------------------- July 11, 2006 (appx 3:15pm EST) Still working on residual history and plotting. Primarily working on "viewResid_Callback" function in "AMGLab_GUI.m". Fixed undefined Callbacks in the GUI are throwing exceptions on mouse over events -Cycle Options Menu, Cycle Type Submenu Added option for user to turn on/off performance profiler More changes we thought of today... Consolidate problem size menu into Poisson submenu (problem source) ------------------------------------------------------------------------------- July 12, 2006 (appx 3:30pm EST) More changes we thought of today... Migrate Settings Menu to Main AMGLab Console Add support for logging performance and settings history, switchable Fixed the "Finite DIFference" type-o in Problem Source menu Residual plotting now works in a way which makes sense and the plots can be zoomed. Added iteration count history and feature to graph iteration history. The software is now ready to collect numerical experiment data, which we will do next. ------------------------------------------------------------------------------- July 17, 2006 (appx 5:00pm EST) I am working on generating finite element Poisson example problems. This code is located in "amg_example1.m". In lines 31-45 there is a loop which generates the linear triangular elements. I have verified that the connectivities are correct for each element. However, I am having trouble assigning the values in the 3x3 edge matrices for the points to make this represent a Poisson problem. ------------------------------------------------------------------------------- July 19, 2006 (appx 12:30pm EST) At the direction of Dr. Gundolf Haase I have modified the code in "amg_example1.m" lines 31-45. The result is now a matrix closer to Poisson 5-point stencil but it is not exactly the same. I think this is due to the accumulation of the matrix from many linear triangular elements. The resulting stiffness matrix has good solvable characteristics and a good approximation is easily obtained by the provided solution schemes. ------------------------------------------------------------------------------- July 26, 2006 (appx 3:45pm EST) Max iterations for smoothers is now user selectable. As of the end of this session I have frozen AMGLab for a new release version (1.2)

近期下载者：

相关文件：

评论：[我要评论] [举报此文件]

收藏者：