[q-lang-users] Floating point arithmetics

[<jir...@bl...>]

Hello Q users,

In order to calculate some linear regressions I ported a few routines 
from the BLAS library to the Q (solving system of linear equations, 
matrix inversion). Experimenting with this routines I found that 
calculations are quite imprecise. E.g. inversion of Hilbert matrix 
(which behaves very badly in this respect) of size of 5 leads to an 
error of approximately 1e-5. Realizing the same algorithm in Erlang 
leads to an error of circa 1e-13.

Therefore I would like to ask how Q does represent floating point 
numbers and whether there is some workaround of that.

Many thanks,

Jiri Spitz