Parallel Computer Architecture

The two key parametric concerns in designing parallel computer architecture are:

1) executing many numbers of instructions concurrently, and

2) increasing the efficiency of processors. There are several methods by that instructions may be executed concurrently and parallel architectures are relied on these techniques of executing instructions in parallel.

Pipelining is the classical and effective method to boost parallelism where various stages execute repeated functions on several operands. Vector processing is the logical or arithmetic computation applied on vectors while in scalar processing just one data item or a pair of data items is processed. Parallel architectures too have been developed based on associative memory organizations. Another concept of improving the processor's speediness by having many instructions for each cycle is termed as Superscalar processing. Multithreading for escalating processor utilization has also been used in 'parallel computer architecture'.