so if you are in slurm
srun -n1 -o t.txt matlab -nodesktop -nosplash -nodisplay -r “temp4(29)” &

t.txt <— will get all the output

after -r you can give all the commands as the form of script i.e. ” “

in above case temp4 was matlab file function, and 29 was input

& so that it can run in background

-n1 only one measly node

But then you can run multiple of such commands and each one gets node

Don’t forget to open the matlabpool in the file you are running if you are using the parallel commands.

function mat = getPatch(srcMat, rect)
%function getPatch(srcMat, rect)
%rect is [x y width height]

stX = max(rect(1),1);stY = max(rect(2),1);
endX = min(rect(1)+rect(3)-1, size(srcMat,1));endY = min(rect(2)+rect(4)-1, size(srcMat,2));
mat = srcMat(stY:endY, stX:endX);

poked around to find what’s happening, firewall was fine and I was on the administrator account, but it appears one has to set the application it-self as “run as administrator”

i.e. go to the XWin file, right click, get the properties, go to advanced and check  ”Run as administrator”.

http://www8.cs.umu.se/~isak/snippets/mdalloc.c

some other code snippets are http://www8.cs.umu.se/~isak/snippets/

http://www.vlfeat.org/index.html

function D = calculateDistMatrix( points1, points2, f)
%D = calculateDistMatrix( points1, points2)
%points1: is m by n matrix, where n is number of points
if nargin <3
    f = 0;
end
totalPoints1 = size(points1,2);
totalPoints2 = size(points2,2);
if f == 0
D= zeros(totalPoints1, totalPoints2);
else
D = sparse(totalPoints1, totalPoints2);
end
for i=1:totalPoints1
    D(i,:) = sqrt(sum((points2-repmat(points1(:,i), 1,totalPoints2)).^2));
end

Only headache is I am having problem in compiling with the VC. They don’t have any dsp file for compilation of code in Windows, so don’t know which libaries to create. Have to use cygwin.

Comment: Ctrl+K+C

Un-Comment: Ctrl+K+U

Download it, run it.

Follow these two links for installation help

• Deploying applications with Intel® IPP DLLs
• IPP with OpenCV highGUI + VS 2008 express ; as the name shows is more about making OpenCV and IPP work together. It even has the test code in it.

Before we move ahead, let’s look at the new version of OpenCV. It’s install guide is here. They were recommending CMake, so tried it. Because we are using the IPP here also so carefully have to give the bin address.

The IPP it appears, by searching online,  not to work properly with OpenCV’s new version. Had actually some problem, but it was I think because IPP libraries access the OpenCv libraries and these libraries have their names changed. So copy pasted them and changed new names to old names. Compiled and viola it’s linking.

Well while putting it here, it looks everything is working fine but it took some time before i was able to solve that.

Some lessons actually commonsense things which one forgets the moment you need it most.

1) Every time you add the Environment variable, restart Visual Studio

2) Bad idea to copy paste your dll’s, set their location path in the Environment Variable path.