RUI: Testing and Enhancing a Prototype Program Fusion Engine

9900510贝茨，Patricia模块化程序结构提供了许多好处，包括提高了软件组件的可靠性、清晰度和可维护性。不幸的是，模块化程序的效率通常低于非模块化程序。热融合是一种提高模块化构建程序的效率的技术，它通过消除数据结构来提高程序的效率，这些数据结构的唯一目的是将较小的程序组件“粘合”成较大的程序组件。最近开发了一种全自动热融合引擎，该原型成功地从受限但普遍存在的函数式程序类中消除了这种数据结构。本研究致力于扩展现有的引擎，以适应更灵活、更广泛适用的热融合启发式算法。它还涉及对扩展原型的全面测试，以便确定实施的温融合技术在实践中成功和失败的条件；这些信息将用于量化可归因于温融合的计划改进，以及开发更通用的融合技术。研究的最终目标是充分增强现有引擎，以便生成的系统适合并入最先进的函数式语言编译器。

9900510 Bates, PatriciaModular program construction offers many benefits, including increased reliability, clarity, and maintainability of software components. Unfortunately, modular programs are typically less efficient than their nonmodular counterparts. Warm fusion is a technique for improving the efficiency of modularly constructed programs by eliminating from them data structures whose sole purpose is to "glue" together smaller program components into larger ones. A fully automatic warm fusion engine has recently been developed, and this prototype successfully eliminates such data structures from a restricted, but pervasive, class of functional programs. This research focuses on extending the existing engine to accommodate more flexible and more widely applicable warm fusion heuristics. It also involves comprehensive testing of the extended prototype so that conditions under which the implemented warm fusion techniques succeed and fail in practice can be determined; this information will be used to quantify program improvement attributable to warm fusion, as well as to develop more general fusion techniques. The ultimate goal of the research is to sufficiently enhance the existing engine so that the resulting system is suitable for incorporation into state-of-the-art functional language compilers.