numerical Module Reference

platform-dependant numerical resolution More...

Variables
real, parameter	machprec = 1.11e-16
	machine precision: 0.5*epsilon(one) for 64 bit double precision
real, parameter	vsmall = 100*machprec
	very small number
real, parameter	small = 10000*machprec
	small number
real, parameter	sqrtmachprec = sqrt(machprec)
	square root of machine precision
real, parameter	logtolerance = 1.0e-32
	this is used to avoid taking alog10(zero); see e.g. dforce;

Detailed Description

platform-dependant numerical resolution