#include "common.h"
#include <float.h>
Go to the source code of this file.
Functions | |
double | bn_epsilon () |
float | bn_epsilonf () |
double | bn_dbl_min () |
double | bn_dbl_max () |
double | bn_flt_min () |
double | bn_flt_max () |
double | bn_ulp (double val) |
Routines useful for performing comparisons and dynamically calculating floating point limits including the Unit in the Last Place (ULP).
In this context, ULP is the distance to the next normalized floating point value larger that a given input value.
TODO: handle NaN, +-Inf, underflow, overflow, non-IEEE, float.h
This file is completely in flux, incomplete, limited, and subject to drastic changes. Do NOT use it for anything.
It also assumes an IEEE 754 compliant floating point representation.
Definition in file ulp.c.