A study of irreducibility in C programs

Stanier, James; Watson, D. M. S.

doi:10.1002/spe.1059

Cited by 16 publications

(7 citation statements)

References 35 publications

(46 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Figure 7 shows its improvement over SCFR in terms of branch statements. Static branch counts are not directly proportional to program runtime, but we regard them as indicative of overheads incurred by other methods [Janssen and Corporaal 1997;Unger and Mueller 2002;Stanier and Watson 2011]. We note that graphs generated by SCFR greatly resemble results produced by Johnson's algorithm [Johnson 2004], suggesting that their overheads are comparable.…”

Section: Key Observationsmentioning

confidence: 94%

“…Janssen and Corporaal [1997] and Unger and Mueller [2002] use controlled node splitting to transform irreducible CFGs to reducible ones. Although irreducible control flow is extremely rare in practice [Stanier and Watson 2011] and their results are encouraging in terms of code bloat, it is proven that node splitting results in exponential blowup for particular CFGs [Carter et al 2003]. …”

Section: Control Flow Restructuringmentioning

confidence: 99%

See 1 more Smart Citation

Perfect Reconstructability of Control Flow from Demand Dependence Graphs

Bahmann

Reißmann

Jahre

et al. 2015

ACM Trans. Archit. Code Optim.

View full text Add to dashboard Cite

Demand-based dependence graphs (DDGs), such as the (Regionalized) Value State Dependence Graph ((R)VSDG), are intermediate representations (IRs) well suited for a wide range of program transformations. They explicitly model the flow of data and state, and only implicitly represent a restricted form of control flow. These features make DDGs especially suitable for automatic parallelization and vectorization, but cannot be leveraged by practical compilers without efficient construction and destruction algorithms. Construction algorithms remodel the arbitrarily complex control flow of a procedure to make it amenable to DDG representation, whereas destruction algorithms reestablish control flow for generating efficient object code. Existing literature presents solutions to both problems, but these impose structural constraints on the generatable control flow, and omit qualitative evaluation.The key contribution of this article is to show that there is no intrinsic structural limitation in the control flow directly extractable from RVSDGs. This fundamental result originates from an interpretation of loop repetition and decision predicates as computed continuations, leading to the introduction of the predicate continuation normal form. We provide an algorithm for constructing RVSDGs in predicate continuation form, and propose a novel destruction algorithm for RVSDGs in this form. Our destruction algorithm can generate arbitrarily complex control flow; we show this by proving that the original CFG an RVSDG was derived from can, apart from overspecific detail, be reconstructed perfectly. Additionally, we prove termination and correctness of these algorithms. Furthermore, we empirically evaluate the performance, the representational overhead at compile time, and the reduction in branch instructions compared to existing solutions. In contrast to previous work, our algorithms impose no additional overhead on the control flow of the produced object code. To our knowledge, this is the first scheme that allows the original control flow of a procedure to be recovered from a DDG representation.

show abstract

Section: Key Observationsmentioning

confidence: 94%

Section: Control Flow Restructuringmentioning

confidence: 99%

Perfect Reconstructability of Control Flow from Demand Dependence Graphs

Bahmann

Reißmann

Jahre

et al. 2015

ACM Trans. Archit. Code Optim.

View full text Add to dashboard Cite

show abstract

“…For simplicity of presentation, we assume that every client CFA contains only one call to the library, and that every client CFA is reducible. We relax the first assumption in Sec 4, and note that the second assumption is innocuous since irreducible control-flow is rare in practice [33]. For the remainder of this section, let be the old version of the library CFA, be the new version of the library CFA, and be the client CFA with a single library call on an edge .…”

Section: Clever At a High Levelmentioning

confidence: 99%

Scaling client-specific equivalence checking via impact boundary search

Feng

Mora²,

Hui

et al. 2020

Proceedings of the 35th IEEE/ACM International Conference on Automated Software Engineering

View full text Add to dashboard Cite

Client-specific equivalence checking (CSEC) is a technique proposed previously to perform impact analysis of changes to downstream components (libraries) from the perspective of an unchanged system (client). Existing analysis techniques, whether general (regression verification, equivalence checking) or special-purpose, when applied to CSEC, either require users to provide specifications, or do not scale. We propose a novel solution to the CSEC problem, called 2clever, that is based on searching the control-flow of a program for impact boundaries. We evaluate a prototype implementation of 2clever on a comprehensive set of benchmarks and conclude that our prototype performs well compared to the state-of-the-art.

show abstract

“…This choice is justified by the fact that irreducible control flow is very rare in practice (e.g. see the study in [25]). For analyzing irreducible programs we propose to use program transformations that make the program reducible; we do not elaborate this idea further due to lack of space.…”

Section: Lossy Vasss and Basic Definitionsmentioning

confidence: 99%

A Simple and Scalable Static Analysis for Bound Analysis and Amortized Complexity Analysis

Sinn

Zuleger

Veith

2014

Computer Aided Verification

View full text Add to dashboard Cite

Abstract. We present the first scalable bound analysis that achieves amortized complexity analysis. In contrast to earlier work, our bound analysis is not based on general purpose reasoners such as abstract interpreters, software model checkers or computer algebra tools. Rather, we derive bounds directly from abstract program models, which we obtain from programs by comparatively simple invariant generation and symbolic execution techniques. As a result, we obtain an analysis that is more predictable and more scalable than earlier approaches. We demonstrate by a thorough experimental evaluation that our analysis is fast and at the same time able to compute bounds for challenging loops in a large real-world benchmark. Technically, our approach is based on lossy vector addition systems (VASS). Our bound analysis first computes a lexicographic ranking function that proves the termination of a VASS, and then derives a bound from this ranking function. Our methodology achieves amortized analysis based on a new insight how lexicographic ranking functions can be used for bound analysis.

show abstract

A study of irreducibility in C programs

Cited by 16 publications

References 35 publications

Perfect Reconstructability of Control Flow from Demand Dependence Graphs

Perfect Reconstructability of Control Flow from Demand Dependence Graphs

Scaling client-specific equivalence checking via impact boundary search

A Simple and Scalable Static Analysis for Bound Analysis and Amortized Complexity Analysis

Contact Info

Product

Resources

About