An intermediate language interpreter for the C-Based Mumps programming language

The C-Based Mumps programming language is an implementation of ANSI standard Mumps. It is a fully interpreted high-level programming language oriented towards string manipulation and database applications.

This project studied the potential for execution speed improvement by the elimination of parsing overhead while maintaining complete language functionality. This presents difficulties because the Mumps language definition includes dynamic code formation and execution which precludes full compiling.

Compilation difficulties are overcome by using an intermediate language. Mumps is translated into an intermediate language, which fan then be interpreted without additional parsing. An intermediate language program is interpreted by a special interpreter that is integrated into the full Mumps interpreter. Dynamic situations are handled by communication between the two interpreters.

The major focus of the project was the design of the intermediate language and the implementation of its interpreter. In addition, an intermediate language translator was implemented by modifying a portion of the C-Based interpreter to generate intermediate code.

Project results show that the intermediate language technique is capable of maintaining complete Mumps language functionality. Execution speed improvements of from 3% to 45% percent (does not include translation time) were measured. The amount of improvement was directly related to the amount of underlying parsing involved. It was also found that Mumps execution time is heavily dominated by string manipulations and symbol table lookup.