In scalar processors, only one instruction is implemented per cycle. That means only single instruction is issued per cycle and only single instruction is completed. But the speed of the processor can be improved in scalar pipeline processor if only multiple instructions instead of one are issued per cycle. This thought of improving the processor's speed by having multiple instructions per cycle is called as Superscalar processing. In superscalar processing multiple results are generated per cycle and multiple instructions are issued per cycle. Therefore, the basic idea of superscalar processor is to have extra instruction level parallelism.