Analysis of compiler optimization techniques by using feature mining technique

Different compilers offer different kind and different level of optimization techniques. These days, scale of applications is increasing rapidly. This increase in scale demands more complex applications to be developed. This results in imbalance of resource capability and application demand. To take maximum of existing resources, compiler optimization techniques are used. There are existing compiler optimization techniques for single core processors. But use of these optimization techniques requires expert level knowledge. Optimization of Compiler means minimization or maximization or various objectives. These objectives may be latency, size, power requirement etc. Importance of these objectives may vary from one application to other. This paper illustrates analysis of existing optimization techniques, which are available for single core processor systems. The speed-up analysis of different applications is performed and architecture level features are extracted. Analysis graphs are generated at the end to show effect of various program features on applicability of optimization techniques.