< Terug naar vorige pagina
Publicatie
A Parallel Worklist Algorithm and Its Exploration Heuristics for Static Modular Analyses
Tijdschriftbijdrage - Tijdschriftartikel
One way to speed up static program analysis is to make use of today's multi-core CPUs by parallelising the analysis.
Existing work on parallel analysis usually targets traditional data-flow analyses for static, first-order languages such as C.
Less attention has been given so far to the parallelisation of more general analyses that can also target dynamic, higher-order languages such as JavaScript.
These are significantly more challenging to parallelise, as dependencies between analysis results are only discovered during the analysis itself.
State-of-the-art parallel analyses for such languages are therefore usually limited, both in their applicability and performance gains.
In this work, we propose the parallelisation of modular analyses.
Modular analyses compute different parts of the analysis in isolation of one another, and therefore offer inherent opportunities for parallelisation that have not been explored so far.
In addition, they can be used to develop a general class of analysers for dynamic, higher-order languages.
We present a parallel variant of the worklist algorithm that is used to drive such modular analyses.
To further speed up its convergence, we show how this algorithm can exploit the monotonicity of the analysis.
Existing modular analyses can be parallelised without additional effort by instead employing this parallel worklist algorithm.
We demonstrate this for ModF, an inter-procedural modular analysis, and for ModConc, an inter-process modular analysis.
For ModConc, we reveal an additional opportunity to exploit even more parallelism in the analysis:
analyses of individual ModConc components can themselves be parallel, resulting in a doubly-parallel exploration.
Finally, we present several heuristics for the exploration order of the analysis and discuss how they can impact its performance.
The parallel worklist algorithm and the exploration heuristics are implemented for and integrated into MAF, a framework for modular program analysis.
On a set of Scheme benchmarks for ModF, we observe speedups between 3x and 8x when using 4 workers, and speedups between 8x and 32x when using 16 workers, with a maximum speedup of 333x using 128 workers.
For ModConc, we achieve a maximum speedup of 37x with 32 workers.
We observe that on a ModF analysis, among 11 exploration heuristics, the heuristics prioritising either components with smaller environments or with less dependencies result in consistent speedups that can reach 20x those of a random exploration strategy.
We find a clear correlation between the mean number of dependencies in a program and the speedup obtained by this heuristic.
Existing work on parallel analysis usually targets traditional data-flow analyses for static, first-order languages such as C.
Less attention has been given so far to the parallelisation of more general analyses that can also target dynamic, higher-order languages such as JavaScript.
These are significantly more challenging to parallelise, as dependencies between analysis results are only discovered during the analysis itself.
State-of-the-art parallel analyses for such languages are therefore usually limited, both in their applicability and performance gains.
In this work, we propose the parallelisation of modular analyses.
Modular analyses compute different parts of the analysis in isolation of one another, and therefore offer inherent opportunities for parallelisation that have not been explored so far.
In addition, they can be used to develop a general class of analysers for dynamic, higher-order languages.
We present a parallel variant of the worklist algorithm that is used to drive such modular analyses.
To further speed up its convergence, we show how this algorithm can exploit the monotonicity of the analysis.
Existing modular analyses can be parallelised without additional effort by instead employing this parallel worklist algorithm.
We demonstrate this for ModF, an inter-procedural modular analysis, and for ModConc, an inter-process modular analysis.
For ModConc, we reveal an additional opportunity to exploit even more parallelism in the analysis:
analyses of individual ModConc components can themselves be parallel, resulting in a doubly-parallel exploration.
Finally, we present several heuristics for the exploration order of the analysis and discuss how they can impact its performance.
The parallel worklist algorithm and the exploration heuristics are implemented for and integrated into MAF, a framework for modular program analysis.
On a set of Scheme benchmarks for ModF, we observe speedups between 3x and 8x when using 4 workers, and speedups between 8x and 32x when using 16 workers, with a maximum speedup of 333x using 128 workers.
For ModConc, we achieve a maximum speedup of 37x with 32 workers.
We observe that on a ModF analysis, among 11 exploration heuristics, the heuristics prioritising either components with smaller environments or with less dependencies result in consistent speedups that can reach 20x those of a random exploration strategy.
We find a clear correlation between the mean number of dependencies in a program and the speedup obtained by this heuristic.
Tijdschrift: Journal of Systems and Software
ISSN: 0164-1212
Volume: 181
Jaar van publicatie:2021
BOF-keylabel:ja
IOF-keylabel:ja
BOF-publication weight:1
Auteurs:Regional
Authors from:Higher Education
Toegankelijkheid:Open