Following up from this older question, I understand that calculation of determinants for integer-valued matrices is possible with polynomial scaling. However, I have been unable to locate any resources that describe computational complexities for matrices of floating-point numbers. I am particularly interested in the calculations of slater determinants using double precision, and would therefore like to ask the following question:
What is the lower (or upper) complexity bound for calculating the determinant of an $n$x$n$ matrix containing finite-precision floating-point numbers (e.g. single or double precision)? What about complex entries using floating point arithmetic?