This paper presents a general framework for deriving demanddriven algorithms for interprocedural data flow analysis of imperative programs. The goal of demand-driven analysis is to reduce the time and/or space overhead of conventional exhaustive analysis by avoiding the collection of information that is not needed. In our framework, a demand for data flow information is modeled as a set of data flow queries. The derived demand-driven algorithms find responses to these queries through a partial reversal of the respective data flow analysis. Depending on whether minimizing time or space is of primary concern, result caching may be incorporated in the derived algorithm. Our framework is applicable to interprocedural data flow problems with a finite domain set. If the problem's flow functions are distributive, the derived demand algorithms provide as precise information as the corresponding exhaustive analysis. For problems with monotone but non-distributive flow functions the provided dat...

Building efficient tools for understanding large software systems is difficult. Many existing program understanding tools build control-flow and data-flow representations of the program a priori, and therefore may require prohibitive space and time when analyzing large systems. Since much of these representations may be unused during an analysis, we construct representations on demand, not in advance. Furthermore, some representations, such as the abstract syntax tree, may be used infrequently during an analysis. We discard these representations and recompute them as needed, reducing the overall space required. Finally, we permit the user to selectively trade-off time for precision and to customize the termination of these costly analyses in order to provide finer user control. We revised the traditional software architecture for compilers to provide these features without unnecessarily complicating the analyses themselves.

this article, we present a general framework for developing demand-driven interprocedural data flow analyzers and report our experience in evaluating the performance of this approach. A demand for data flow information is modeled as a set of queries. The framework includes a generic demand-driven algorithm that determines the response to a query by iteratively applying a system of query propagation rules. The propagation rules yield precise responses for the class of distributive finite data flow problems. We also describe a two-phase framework variation to accurately handle nondistributive problems. A performance evaluation of our demand-driven approach is presented for two data flow problems, namely, reaching-definitions and copy constant propagation. Our experiments show that demand-driven analysis performs well in practice, reducing both time and space requirements when compared with exhaustive analysis.

Reducing communication cost is crucial to achieving good performance on scalable parallel machines. This paper presents a new compiler algorithm for global analysis and optimization of communication in data-parallel programs. Our algorithm is distinct from existing approaches in that rather than handling loop-nests and array references one by one, it considers all communication in a procedure and their interactions under different placements before making a final decision on the placement of any communication. It exploits the flexibility resulting from this advanced analysis to eliminate redundancy, reduce the number of messages, and reduce contention for cache and communication buffers, all in a unified framework. In contrast, single loop-nest analysis often retains redundant communication, and more aggressive dataflow analysis on array sections can generate too many messages or cache and buffer contention. The algorithm has been implemented in the IBM pHPF compiler for High Performan...

. This paper addresses the problems of representing aliases and indirect memory operations in SSA form. We propose a method that prevents explosion in the number of SSA variable versions in the presence of aliases. We also present a technique that allows indirect memory operations to be globally commonized. The result is a precise and compact SSA representation based on global value numbering, called HSSA, that uniformly handles both scalar variables and indirect memory operations. We discuss the capabilities of the HSSA representation and present measurements that show the effects of implementing our techniques in a production global optimizer. Keywords. Aliasing, Factoring dependences, Hash tables, Indirect memory operations, Program representation, Static single assignment, Value numbering. 1 Introduction The Static Single Assignment (SSA) form [CFR+91] is a popular and efficient representation for performing analyses and optimizations involving scalar variables. Effecti...

We present a programming language called TCEL (Time-Constrained Event Language), whose semantics is based on time-constrained relationships between observable events. Such a semantics infers only those timing constraints necessary to achieve real-time correctness, without over-constraining the system. Moreover, an optimizing compiler can exploit this looser semantics to help tune the code, so that its worst-case execution time is consistent with its real-time requirements. In this paper we describe such a transformation system, which works in two phases. First the TCEL source code is translated into an intermediate representation. Then an instructionscheduling algorithm rearranges selected unobservable operations, and synthesizes tasks guaranteed to respect the original event-based constraints. Keywords: Real-time, programming languages, compiler optimization, code scheduling, single-static assignment, timing analysis, trace scheduling, code motion. This research is supported in par...

In recent years, the trend in program representations for imperative programs has been to make them more functional, or to make them more sparse. However, new sparse representations have been nonfunctional, and new functional representations have not been sparse in the presence of pointer operations. In this paper, we present a functional representation that is sparse even in the presence of pointer operations. Conventionally, a store is represented in a functional program representation by a single object -- typically a mapping from locations to values. We show how such a store object may be fragmented into several objects, each representing part of the store. The result is a sparser representation, which has not only the usual benefit of directly linking producers to consumers, but which also for static program analysis often leads to smaller domains of abstract values for store objects. Store fragmentation corresponds to assignment factored SSA form (a factorization of SSA...

Software is difficult and costly to modify correctly. Automating tiresome mechanical tasks such as program restructuring is one approach to reducing the burden of software maintenance. Several restructuring tools have been proposed and prototyped, all centered on the concept of meaning-preserving transformations similar in spirit to compiler optimizations. Like optimizing compilers, these tools rely on static analysis to reason about the correctness of program changes. However, the cost (in both time and space) of static analysis serves as the limiting factor for transformation tools, resulting in slow, complex tool designs that scale poorly for use on large systems. To reduce these costs, this thesis proposes efficient, demand-driven flow analysis techniques as an alternate to traditional, compiler-based methods. These techniques operate directly on the abstract syntax tree (AST), the data structure most appropriate for use in a source-to-source tool architecture. By eliminating the need for other program representations such as the standard control flow graph (CFG) or program dependence graph (PDG), this approach greatly simplifies program modification. A key contribution of this work is the idea of virtual control flow, a method for computing the control successors or predecessors of individual AST expressions on demand. This method handles all types of structured and unstructured jumps found in an imperative programming language such as C. Virtual control flow couples well with demand-driven data flow analysis to minimize the cost of determining semantic information. To conservatively estimate data flow relationships, the effects of aliasing between memory locations can be inexpensively approximated using flow-insensitive points-to analysis based on type inference. These techniques were implemented in a prototype tool called Cstructure to support a simple restructuring transformation for reordering program statements. To check that this transformation does not change the program's behavior requires syntax, control flow and data dependence analysis. Experimental results on three programs ranging in size from 72,000 to 213,000 lines of code demonstrate the performance advantages of such aggressive demand-driven approaches. For the largest program, gcc, check times for the average statement were 50 milliseconds on a desktop workstation.

Many abstract value domains such as intervals, bitwise, constants, and value-sets have been developed to support dataflow analysis. Different domains offer alternative tradeoffs between analysis speed and precision. Furthermore, some domains are a better match for certain kinds of code than others. This paper presents the design and implementation of cXprop, an analysis and transformation tool for C that implements “conditional X propagation,” a generalization of the well-known conditional constant propagation algorithm where X is an abstract value domain supplied by the user. cXprop is interprocedural, context-insensitive, and achieves reasonable precision on pointer-rich codes. We have applied cXprop to sensor network programs running on TinyOS, in order to reduce code size through interprocedural dead code elimination, and to find limited-bitwidth global variables. Our analysis of global variables is supported by a novel concurrency model for interruptdriven software. cXprop reduces TinyOS application code size by an average of 9.2 % and predicts an average data size reduction of 8.2 % through RAM compression.

Existing methods to handle pointer variables during data-flow analyses can make such analyses inefficient both in time and space because the data-flow analyses must store and propagate large sets of data facts that are introduced by dereferences of pointer variable. This paper presents equivalence analysis, a general technique to improve the efficiency of data-flow analyses in the presence of pointers. The technique identifies equivalence relations among the memory locations accessed by a procedure and ensures that two equivalent memory locations share the same set of data facts in a procedure and in the procedures that are called by that procedure. Thus, a data-flow analysis needs to compute the data-flow information only for a representative memory location in an equivalence class. The data-flow information for other memory locations in the equivalence class can be derived from that of the representative memory location. Our empirical studies indicate that equivalence analysis may ef...