Software engineers are faced with a dilemma. They want to write general and wellstructured programs that are flexible and easy to maintain. On the other hand, generality has a price: efficiency. A specialized program solving a particular problem is often significantly faster than a general program. However, the development of specialized software is time-consuming, and is likely to exceed the production of today’s programmers. New techniques are required to solve this so-called software crisis. Partial evaluation is a program specialization technique that reconciles the benefits of generality with efficiency. This thesis presents an automatic partial evaluator for the Ansi C programming language. The content of this thesis is analysis and transformation of C programs. We develop several analyses that support the transformation of a program into its generating extension. A generating extension is a program that produces specialized programs when executed on parts of the input. The thesis contains the following main results.

Specialization has been recognized as a powerful technique for optimizing operating systems. However, specialization has not been broadly applied beyond the research community because current techniques, based on manual specialization, are time-consuming and error-prone. The goal of the work described in this paper is to help operating system tuners perform specialization more easily. We have built a specialization toolkit that assists the major tasks of specializing operating systems. We demonstrate the effectiveness of the toolkit by applying it to three diverse operating system components. We show that using tools to assist specialization enables significant performance optimizations without errorprone manual modifications. Our experience with the toolkit suggests new ways of designing systems that combine high performance and clean structure. 1

We investigate the application of partial evaluation to numerically oriented computation in scientific and engineering applications. We present our results using the Fast Fourier Transformation, the N-body attraction problem, and the cubic splines interpolation as examples. All programs are written in Fortran 77, specialized using our Fortran partial evaluator, and compiled into executable machine code using a commercial Fortran compiler. The results demonstrate that existing partial evaluation technology is strong enough to improve the efficiency of a large class of numerical programs. However, using partial evaluation as a development tool in the ‘real world’ still remains a challenging problem.

Program specialization can divide a computation into several computation stages. We present the key ingredient of our approach to multi-level specialization: an accurate and fast multi-level binding-time analysis. Three e#cient program analyses for higher-order, functional languages are presented which are based on constraint systems and run almost-linear in the size of the analyzed programs. The three constraint normalizations have been proven correct (soundness, completeness, termination, existence of best solution). The analyses have all been implemented for a substantial, higher-order subset of Scheme. Experiments with widely-available example programs confirm the excellent run-time behavior of the normalization algorithms.

Introduction: What is partial evaluation? Partial evaluation is a technique to partially execute a program, when only some of its input data are available. Consider a program p requiring two inputs, x 1 and x 2 . When specific values d 1 and d 2 are given for the two inputs, we can run the program, producing a result. When only one input value d 1 is given, we cannot run p, but can partially evaluate it, producing a version p d1 of p specialized for the case where x 1 = d 1 . Partial evaluation is an instance of program specialization, and the specialized version p d1 of p is called a residual program. For an example, consider the following C function p

In this paper, we present an abstraction for heap-allocated storage, called the recency-abstraction, that allows abstract-interpretation algorithms to recover some non-trivial information for heap-allocated data objects. As an application of the recency-abstraction, we show how it can resolve virtual-function calls in stripped executables (i.e., executables from which debugging information has been removed). This approach succeeded in resolving 55% of virtual-function call-sites, whereas previous tools for analyzing executables fail to resolve any of the virtual-function call-sites.

Object-oriented programming facilitates the development of generic software, but at a significant cost in terms of performance. We apply partial evaluation to object-oriented programs, to automatically map generic software into specific implementations. In this paper we give a concise, formal description of a simple partial evaluator for a minimal object-oriented language, and give directions for extending this partial evaluator to handle realistic programs. 1

Abstract not found

This paper reviews the principles behind the paradigm of "abstract interpretation via 3-valued logic," discusses recent work to extend the approach, and summarizes ongoing research aimed at overcoming remaining limitations on the ability to create programanalysis algorithms fully automatically.

Abstract. We have developed and implemented a partial evaluator for a subset of Fortran 77. A partial evaluator is a tool for program transformation which takes as input a general program and a part of its input, and produces as output a specialized program. The goal is efficiency: a specialized program often runs an order of magnitude faster than the general program. The partial evaluator is based on the off-line approach and uses a binding-time analysis prior to the specialization phase. The source language includes multi-dimensional arrays, procedures and functions, as well as global storage. The system is presented and experimental results are given. 1