function []=hw4_2_8()
 
% perform the Cholesky decomposition on a matrix as a sparse matrix and as a full
% matrix with eigenvalues r.
r=[3:2:21];
A=sprandsym(10,0.2,r);

% try timing instead of flops
tic
C=chol(A);
t1=toc;
disp('sparse time')
disp(t1)

% convert to a full matrix and retime
fA=full(A);
tic
fC=chol(fA);
t2=toc;
disp('full time')
disp(t2)

% the only reason the sparse time is longer is because the sparse computation is
% done first, so when the full decomposition is computed it already "knows" the
% answer