I have a very large sparse matrix, Amat (192611x192611), that has a density of 6.7190e-05 . When I am trying to use the backslash command with this matrix (cnew =Amat/rhs, where rhs is a column vector (192611 x 1) with small values, around 1e-13) on matlab 2019a, I get an error that says ''Warning: Matrix is singular to working precision." and the output array is full of NaN. However, when I run the same command on matlab 2011b (the code was made in 2011) or matlab 2014a, I get the same error, but I don't get these Nan values. Instead, I get small values, around 1e-14. My questions are the following :
1) Has the backslash command changed between these versions?
2) Are the values that I'm getting with the other versions reliable?