finding inverse of 3D matrix for speed

Hi ..

In a 1D simulation, I have 'n' number of grid points and at each point, some data is stored in 3x3 matrix (A). For simplicity, lets say

tic;
n= 200000;
A = rand(3,3,n);
for i = 1:n
    iA(:,:,i) = inv(A(:,:,i));
end
toc;  Elapsed time is 0.536919 seconds.

With this the code is too slow because of the for loop. If there is any other efficient way anyone knows, pl suggest. The usual way of using "iA = inv(A)" fails as the matrix need to be 2D for using that command.

Thanks.

Answer 1

[James Tursa]

See this small size linear solver by Bruno Luong:

http://www.mathworks.com/matlabcentral/fileexchange/27762-small-size-linear-solver

and you might be interested in this one as well:

http://www.mathworks.com/matlabcentral/fileexchange/24260-multiple-same-size-linear-solver

If you have a C compiler, you could consider this submission by Yuval also:

http://www.mathworks.com/matlabcentral/fileexchange/37515-mmx-multithreaded-matrix-operations-on-n-d-matrices