A Fast and Precise Static Loop Analysis Based on Abstract Interpretation, Program Slicing and Polytope Models

Lokuciejewski, Paul; Cordes, Daniel; Falk, Heiko; Marwedel, Peter

doi:10.1109/cgo.2009.17

Cited by 39 publications

(8 citation statements)

References 26 publications

(26 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Pointer-based array traversals are analyzed and transformed to closed form array expressions in [15]. Static analysis is employed in [9] to determine loop iteration counts using polytope-based loop evaluation and program slicing. A constraint based approach to recognition of reductions is presented in [18], where a wide class of reductions including their loop iterators is recognized in the LLVM framework.…”

Section: Discussionmentioning

confidence: 99%

Generalized profile-guided iterator recognition

Manilov

Vasiladiotis

Franke

2018

Proceedings of the 27th International Conference on Compiler Construction

View full text Add to dashboard Cite

Iterators prescribe the traversal of data structures and determine loop termination, and many loop analyses and transformations require their exact identification. While recognition of iterators is a straight-forward task for affine loops, the situation is different for loops iterating over dynamic data structures or involving control flow dependent computations to determine the next data element. In this paper we propose a compiler analysis for recognizing loop iterators code for a wide class of loops. We initially develop a static analysis, which is then enhanced with profiling information to support speculative code optimizations. We have prototyped our analysis in the LLVM framework and demonstrate its capabilities using the SPEC CPU2006 benchmarks. Our approach is applicable to all loops and we show that we can recognize iterators in, on average, 88.1% of over 75,000 loops using static analysis alone, and up to 94.9% using additional profiling information. Existing techniques perform substantially worse, especially for C and C++ applications, and cover only 35-44% of the loops. Our analysis enables advanced loop optimizations such as decoupled software pipelining, commutativity analysis and source code rejuvenation for real-world applications, which escape analysis and transformation if loop iterators are not recognized accurately.

show abstract

Section: Discussionmentioning

confidence: 99%

Generalized profile-guided iterator recognition

Manilov

Vasiladiotis

Franke

2018

Proceedings of the 27th International Conference on Compiler Construction

View full text Add to dashboard Cite

show abstract

“…Automated Parallel Conversion arrived later with static analysis as the basis for generating information about the program, such as Flow and Dependence analyses, which provided the impetus for transformations. Analysis techniques were algebraic such as Linear or Polyhedral, and algorithmic such as Trees and Graphs [11][12][13][14][15][16][17][18][19]. While translating high-level languages to an intermediate representation (IR) and transforming and optimizing the IR is the norm, several researchers have tried the source-to-source conversions as a basis for optimizations and parallel conversions [20][21][22][23][24][25][26][27][28][29][30][31].…”

Section: Previous Workmentioning

confidence: 99%

Inherent Parallelism and Speedup Estimation of Sequential Programs

Kalyur¹,

G.S²

2021

AETiC

View full text Add to dashboard Cite

Although several automated Parallel Conversion solutions are available, very few have attempted, to provide proper estimates of the available Inherent Parallelism and expected Parallel Speedup. CALIPER which is the outcome of this research work is a parallel performance estimation technology that can fill this void. High level language structures such as Functions, Loops, Conditions, etc which ease program development, can be a hindrance for effective performance analysis. We refer to these program structures as the Program Shape. As a preparatory step, CALIPER attempts to remove these shape related hindrances, an activity we refer to as Program Shape Flattening. Programs are also characterized by dependences that exist between different instructions and impose an upper limit on the parallel conversion gains. For parallel estimation, we first group instructions that share dependences, and add them to a class we refer to as Dependence Class or Parallel Class. While instructions belonging to a class run sequentially, the classes themselves run in parallel. Parallel runtime, is now the runtime of the class that runs the longest. We report performance estimates of parallel conversion as two metrics. The inherent parallelism in the program is reported, as Maximum Available Parallelism (MAP) and the speedup after conversion as Speedup After Parallelization (SAP).

show abstract

“…The framework provides an interface through the extension of JBCMM to allow its user to insert loop limits. The loop limits to be inserted can be obtained using manual annotations [3] or automated analysis techniques [12]. Currently, we use a default value as the loop limit for all the loops.…”

Section: Loop Augmentationmentioning

confidence: 99%

A Model-Derivation Framework for Software Analysis

Yildiz¹,

Rensink²,

Bockisch³

et al. 2017

Electron. Proc. Theor. Comput. Sci.

View full text Add to dashboard Cite

Model-based verification allows to express behavioral correctness conditions like the validity of execution states, boundaries of variables or timing at a high level of abstraction and affirm that they are satisfied by a software system. However, this requires expressive models which are difficult and cumbersome to create and maintain by hand. This paper presents a framework that automatically derives behavioral models from real-sized Java programs. Our framework builds on the EMF/ECore technology and provides a tool that creates an initial model from Java bytecode, as well as a series of transformations that simplify the model and eventually output a timed-automata model that can be processed by a model checker such as UPPAAL. The framework has the following properties: (1) consistency of models with software, (2) extensibility of the model derivation process, (3) scalability and (4) expressiveness of models. We report several case studies to validate how our framework satisfies these properties.Comment: In Proceedings MARS 2017, arXiv:1703.0581

show abstract

A Fast and Precise Static Loop Analysis Based on Abstract Interpretation, Program Slicing and Polytope Models

Cited by 39 publications

References 26 publications

Generalized profile-guided iterator recognition

Generalized profile-guided iterator recognition

Inherent Parallelism and Speedup Estimation of Sequential Programs

A Model-Derivation Framework for Software Analysis

Contact Info

Product

Resources

About