Function solve_lower 

pub fn solve_lower(l: &[f64], b: &mut [f64], n: usize, m: usize)

Solve L·X = B in place (L lower n×n, B n×m row-major).