matlab_spmd_2010_fsu

matlab_spmd_2010_fsu, "Parallel MATLAB at FSU: SPMD and Message Passing Programming", a lecture on parallel MATLAB, as implemented on the computer cluster at Florida State University's Department of Scientific Computing, on 13 April 2010.

• matlab_spmd_2010_fsu.tex
• matlab_spmd_2010_fsu.pdf

The files used for this presentation include

• cg_it.m, carries out the conjugate gradient iteration;
• cg_spmd.m, uses CG_IT to seek a solution of an eigenvalue problem;
• configuration.png, a menu that displays some properties of a configuration.
• fem2d_heat_steady.png, a surface plot of the solution to a steady heat equation in 2D.
• fem2d_heat_steady_grid.png, a schematic of a very small finite element grid divided among parallel workers.
• filter_rgb.png, illustrates the filtering of noise from an image.
• fsu_logo.pdf, a logo.
• heat_equation.png, illustrates how the grid for the 2D heat equation might be divided into strips, meaning that the 5 point stencil might need information from another strip.
• image_comments.m, a MATLAB function to do image filtering, with comments.
• image_nocomments.m, a MATLAB function to do image filtering, with no comments.
• prime_fun.m, a MATLAB function to seek prime numbers.
• prime_script.m, a MATLAB script to call prime_fun.
• quad_fun.m, a MATLAB function to estimate an integral.
• step000.png, the initial data for the heat equation.
• step010.png, the data for the heat equation after 10 steps.
• step020.png, the data for the heat equation after 20 steps.
• step040.png, the data for the heat equation after 40 steps.
• step080.png, the data for the heat equation after 80 steps.
• surfsup.m, a MATLAB function to filter the surfsup image.
• surfsup.png, an image to be filtered.
• surfsup2_summary.png, compares filtering on the client and in parallel, after messages were used.
• surfsup_filters_spmd.png, an image filtered in parallel, but without messaging.
• surfsup_summary.png, compares filtering on the client and in parallel, with no messages allowed between parallel workers.