A Comparative Analysis of Techniques for Executing Branched Instructions

In general, the majority of instructions bound for processing and executions consist of straight and branched constructs. A straight instruction set does not rely or depend on processes to reach complete execution. Unlike branched (or conditional) instructions, constructs require outcomes of other processes in order to peak its process executions. Previous works intended to first transform all branched instructions types to straight instructions types due to the structure and architecture of underlying compiler and processing elements known as predication process. While instruction-level parallelism (ILP) architecture ensures that many more parallel processes are executed by overlapping of these processes. This paper presents a comparative analysis of predicted and ILP techniques for processes executions of branched instructions constructs. The paper recommends ILP as the best way to deal with branched instructions in concurrent processes/executions because it supports multiple issues, execution speedups, improves memory utilization, and optimization of complier/processing elements based on selected parameters.