Parameterized Construction of Program Representations for Sparse Dataflow Analyses
نویسندگان
چکیده
Data-flow analyses usually associate information with control flow regions. Informally, if these regions are too small, like a point between two consecutive statements, we call the analysis dense. On the other hand, if these regions include many such points, then we call it sparse. This paper presents a systematic method to build program representations that support sparse analyses. To pave the way to this framework we clarify the bibliography about well-known intermediate program representations. We show that our approach, up to parameter choice, subsumes many of these representations, such as the SSA, SSI and e-SSA forms. In particular, our algorithms are faster, simpler and more frugal than the previous techniques used to construct SSI Static Single Information form programs. We produce intermediate representations isomorphic to Choi et al.’s Sparse Evaluation Graphs (SEG) for the family of data-flow problems that can be partitioned per variables. However, contrary to SEGs, we can handle sparsely problems that are not in this family. Key-words: Sparse Data-Flow Analysis, Compiler, Static Single Assignment, Static Single Information, SSA, SSI, Static Single Use, SSU, Iterated Dominance Frontier, Control-Flow Graph ∗ UFMG † ENS Lyon ‡ UFMG § Inria Représentation de programmes pour l’analyse creuse de flots de données: construction paramétrée Résumé : L’analyse de flot de données, associe en général l’information calculée, aux régions de flot de contrôle. Informellement cette analyse est dite dense, si ces régions sont trop petites, i.e. par exemple restreintes aux points de programme situés entre deux instructions. A l’opposé, cette analyse est dite creuse, si ces régions comprennent de nombreux points consécutifs. Cet article présente une méthode de construction systématique d’une représentation de programme qui permet de manière naturelle l’implémentation d’analyses creuses. Cette forme englobe plusieurs forme existante comme la forme SSA, la forme SSI, ou la forme e-SSA. En particulier, l’algorithme présenté est plus rapide, plus simple et moins gourmand que les méthodes existantes de construction de SSI –Static Single Information. Aussi, la représentation ainsi construite se trouve être isomorphe au graphe d’évaluation creux (Sparse Evaluation Graph — SEG in English) de Choi et al. dans le cas particulier ou le problème d’analyse de flot de données peut être partitionné par variable. Cela dit, contrairement aux SEG, l’approche ici décrite n’est pas restreinte à cette famille de problèmes. Mots-clés : Analysis de flot de données, compilateur, forme à assignation unique, SSA, SSI, SSU, frontière de dominance itérée, graphe de flot de contrôle Parameterized Construction of Program Representations for Sparse Dataflow Analyses 3
منابع مشابه
Mapping Parameterized Cyclo-static Dataflow Graphs onto Configurable Hardware
In recent years, parameterized dataflow has evolved as a useful framework for modeling synchronous and cyclo-static graphs in which arbitrary parameters can be changed dynamically. Parameterized dataflow has proven to have significant expressive power for managing dynamics of DSP applications in important ways. However, efficient hardware synthesis techniques for parameterized datafow represent...
متن کاملThe Static Single Information Form
The Static Single Information (SSI) form is a compiler intermediate representation that allows efficient sparse implementations of predicated analysis and backward dataflow algorithms. It possesses several attractive graph-theoretic properties which aid in program analysis. An extension to SSI form, SSI+, is also presented, along with a complete executable abstract semantics for the representat...
متن کاملParameterized Modeling and Scheduling for Dataflow Graphs
Dataflow has proven to be an attractive computational model for programming DSP applications. A restricted version of dataflow, called Synchronous Dataflow (SDF) is particularly well-suited for modeling a large class of signal processing applications, as it offers strong formal properties and compile-time predictability. Efficient techniques have been developed for generating software implement...
متن کاملParameterized Modeling and Scheduling for Dataflow Graphs1
Dataflow has proven to be an attractive computational model for programming DSP applications. A restricted version of dataflow, called Synchronous Dataflow (SDF) is particularly well-suited for modeling a large class of signal processing applications, as it offers strong formal properties and compile-time predictability. Efficient techniques have been developed for generating software implement...
متن کاملHierarchical Mapping Techniques for signal Processing Systems on Parallel Platforms
Title of dissertation: HIERARCHICAL MAPPING TECHNIQUES FOR SIGNAL PROCESSING SYSTEMS ON PARALLEL PLATFORMS Lai-Huei Wang, Doctor of Philosophy, 2014 Dissertation directed by: Professor Shuvra S. Bhattacharyya Department of Electrical and Computer Engineering Dataflow models are widely used for expressing the functionality of digital signal processing (DSP) applications due to their useful featu...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
عنوان ژورنال:
دوره شماره
صفحات -
تاریخ انتشار 2014