The cycle time of a computer is the time required to change the information in a set of registers. This is also sometimes called the state transition time. The register cycle time of a processor is sometimes referred to as the internal cycle time, clock time, or simply cycle time. Main memory cycle time is usually several times the internal cycle time. The internal cycle time may not be of constant value.
There are basically three different types of cycle-timing organization:
1. Synchronous (fixed). In this scheme, all operations are composed of one or more cycles, with the fundamental time quantum being fixed by the design. Such systems are also referred to as clocked, since a master oscillator (or clock) is used to distribute and define these cycles.
2. Synchronous (variable). This is a slight variation of the first scheme; certain long operations are allowed to take multiple cycles without causing a register state transition. In such systems, there may be several different cycle lengths For example, a register-to-register transfer of information cycle might take one cycle, while a register-to-adder and return-to-register cycle would perhaps be two or three cycles.
3. Asynchronous operation. In a completely asynchronous machine, there is no clock or external mechanism that determines a state transition. Rather, the logic of the system is arranged in stages; when the output value of one stage has been stabilized, the logic signals the input at that stage to admit new operands. In current practice, almost all processors use a fixed synchronous cycle as this is required to support a pipe lined implementation. Access to main memory is sometimes implemented as an asynchronous operation.
Dinesh Thakur holds an B.C.A, MCDBA, MCSD certifications. Dinesh authors the hugely popular