Method and computer system for managing threads

A method of managing threads in a processor in which a second thread from a queue having more than two threads is fetched for execution, when an event occurs, blocking a first thread and the management of the threads into consideration the latency time of the event, with the following steps: a) a control logic in a thread execution control, which is a component of a processor that constantly monitors the current, the first thread of execution, and determines whether the latency of the first thread blocking event is smaller than a settable by a programmer in a threshold register number of machine cycles, and thus a short latency is present; b) when an event with a short latency causes the thread execution control such that the control of the execution is temporarily handed over to the second thread and is returned to the first thread when the event is terminated with a short latency.