Hello, all! I have a linear system of real numbers. I copied the example in the tutorial for using LinearSolve and converted my matrix to a float:
my_solve := proc(A::Matrix)
local sz, local_A, B, sol;
sz := Dimension(A);
local_A := Matrix(A, datatype=float);
B := Vector(1...sz, 1);
sol := LinearSolve(local_A, B);
And when it solved, it got the wrong answer!! It would produce a solution that simply didn't work. When I removed the float conversion step, and just used my original matrix, it worked perfectly.
I would love to understand this better... is it because of rounding?
Comments would be lovely!