real.h File Reference

Typedefs
typedef double	real

Variables
const real	REAL_EPSILON = 4096 * std::numeric_limits<real>::epsilon()
	REAL_EPSILON is a multiple of the round-off error associated with real.

Typedef Documentation

typedef double real

It is possible to select double or single precision here:

• To use float, define REAL_IS_FLOAT below
• To use double, do not define REAL_IS_FLOAT

Cytosim is faster in single precision, but the iterative solver used in Meca::solve() (conjugate-gradient) may fail in adverse conditions.

It is safer, and STRONGLY ADVISED therefore, to use double precision, and to not define 'REAL_IS_FLOAT'