higher bound for a

float point error function is

defined in float math✕