Program Optimization Using Abstract State Machines

Abstract

Usually the result code of source code by a compiler is not necessary the best one, and can be improved to run faster or to use less memory. This kind of improvement is done in compiling phase after parsing. Some good techniques in optimization are in folding the constants, elimination of dead code, or improvement of the loops. Here it is considered the runtime overhead and present how can this be improved. The source is specific for objectoriented languages with late binding, where a name of method to be called is bound to method dynamically. It increases the computation time by a cost of traversing the class hierarchy each time a method is called.

Keywords

• optimization
• abstract state machine (evolving algebra)
• class annotation