MATLAB: How to speed up huge matrix mldivide

huge matrixmldividespeed up

How to speed up huge matrix mldivide? I was solved this matrix on single core but it took too long. So, i was used distributed array using parallel computing tool box, but i didn't get speed up. Using distributed array took too long than single core. How to speed up solution.

Best Answer

distributed arrays on a single machine will generally perform worse than normal arrays - this is because the standard MATLAB mldivide is already intrinsically multithreaded.

If you have access to a capable GPU, and your problem fits in the memory of the GPU, this might be a good option. See this example which compares mldivide performance on the CPU and GPU.

Related Solutions

MATLAB: About parfor and spmd speedup

If your matrices are not very big, then the time required to set up workers and communicate data in to them can overwhelm any saving of using multiple workers.

MATLAB: When to use codistributed arrays

Distributed arrays are most useful when you do not have enough memory to store an entire array on a single machine. By distributing chunks of the original array across all the workers in the pool, you can perform operations on the entire array that you previously couldn't even store.

I would suggest that you take a step back and start by working with distributed arrays rather than codistributed arrays. This allows MATLAB to choose a default distribution scheme for you.

A = rand(6400,6400);
matlabpool open 
dA = distributed(A); % let MATLAB pick a default distribution scheme for the 
                     % distributed array, dA
R = chol(dA);        % example of a function that works for 
                     % distributed arrays - no spmd required

With distributed arrays, you can get started without having to worry about what distribution scheme to use. If you are interested, you can always query the distribution scheme that is currently used by a distributed array like so:

% dA is the distributed array created as above
spmd
codistr = getCodistributor(dA) % inside spmd we can "view" dA as a 
                               % codistributed array and access its
                               % distribution scheme
end

Once your code is working correctly with distributed arrays, you can make minor changes to use codistributed arrays with a specific distribution scheme. Changing the default distribution scheme may improve performance, but as Matt J mentions, choosing the most efficient distribution scheme is problem dependent and will depend heavily on the operations that you want to call.

Best Answer

Related Solutions

MATLAB: About parfor and spmd speedup

MATLAB: When to use codistributed arrays

Related Question