underflow in Czech

(computing) A condition in which the value of a computed quantity is smaller than the smallest non-zero value that can be physically stored; usually treated as an error condition.