Superskalär

Superskalär refers to a property of a processor that allows it to execute more than one instruction per clock cycle. This is achieved by having multiple execution units within the processor that can operate in parallel. A superscalar processor can achieve this by fetching, decoding, and executing multiple instructions simultaneously. To identify instructions that can be executed in parallel, the processor employs sophisticated techniques such as instruction-level parallelism (ILP) detection and scheduling.

The core of a superscalar architecture involves multiple pipelines, each capable of executing a different type of instruction. For example, a processor might have separate integer arithmetic units, floating-point units, and load/store units. When the processor fetches a sequence of instructions, it analyzes them to determine which ones can be executed independently and dispatches them to the appropriate available execution unit. This parallelism can significantly boost the overall performance of the processor, especially for computationally intensive tasks.

However, achieving true superscalar execution is complex. Dependencies between instructions, such as one instruction requiring the result of a previous one, can limit parallelism. Processors use various methods to handle these dependencies, including out-of-order execution, where instructions are executed as soon as their operands are ready, rather than in the original program order. Branch prediction is another crucial technique used to keep the pipelines full by guessing the outcome of conditional branches. The effectiveness of a superscalar processor is measured by its degree of superscalarity, which is the average number of instructions it can execute per clock cycle.