Performance-influence models for highly configurable systems

Siegmund, Norbert; Grebhahn, Alexander; Apel, Sven; Kästner, Christian

doi:10.1145/2786805.2786845

Cited by 202 publications

(253 citation statements)

References 31 publications

Supporting

Mentioning

245

Contrasting

Unclassified

Order By: Relevance

“…In prior work, we demonstrated that SPL Conqueror's approach is able to learn accurate performance‐influence models after measuring only a small number of system variants . There, we demonstrated that this even works for complex systems, such as the Java garbage collector.…”

Section: Introductionmentioning

confidence: 83%

“…Knowing the performance of all variants of a configurable system is essential to identify the optimal system variant, to identify specific performance characteristics, such as performance bugs or unknown interactions between configuration options, and, as result, to understand the overall system. To this end, we aim at learning a performance‐influence model of the system that captures the influence of configuration options and their interactions on performance in a human‐readable form . To learn a performance‐influence model, we have to select a representative set of system variants and learn performance characteristics on the basis of the observed performance of this set.…”

Section: Performance‐influence Modelsmentioning

confidence: 99%

See 1 more Smart Citation

Performance‐influence models of multigrid methods: A case study on triangular grids

Grebhahn

Rodrigo

Siegmund

et al. 2017

Concurrency and Computation

Self Cite

View full text Add to dashboard Cite

Summary Multigrid methods are among the most efficient algorithms for solving discretized partial differential equations. Typically, a multigrid system offers various configuration options to tune performance for different applications and hardware platforms. However, knowing the best performing configuration in advance is difficult, because measuring all multigrid system variants is costly. Instead of direct measurements, we use machine learning to predict the performance of the variants. Selecting a representative set of configurations for learning is nontrivial, although, but key to prediction accuracy. We investigate different sampling strategies to determine the tradeoff between accuracy and measurement effort. In a nutshell, we learn a performance‐influence model that captures the influences of configuration options and their interactions on the time to perform a multigrid iteration and relate this to existing domain knowledge. In an experiment on a multigrid system working on triangular grids, we found that combining pair‐wise sampling with the D‐Optimal experimental design for selecting a learning set yields the most accurate predictions. After measuring less than 1 % of all variants, we were able to predict the performance of all variants with an accuracy of 95.9 %. Furthermore, we were able to verify almost all knowledge on the performance behavior of multigrid methods provided by 2 experts.

show abstract

Section: Introductionmentioning

confidence: 83%

Section: Performance‐influence Modelsmentioning

confidence: 99%

Performance‐influence models of multigrid methods: A case study on triangular grids

Grebhahn

Rodrigo

Siegmund

et al. 2017

Concurrency and Computation

Self Cite

View full text Add to dashboard Cite

show abstract

“…The e↵ectiveness of statistical learning techniques and regression methods have been empirically studied. Siegmund et al [29] combined machine-learning and sampling heuristics to compute the individual influences of configuration options and their interactions. The approach has applications in performance-bug detection or configuration optimization.…”

Section: Related Workmentioning

confidence: 99%

“…We use a random strategy for picking values (see Section 3.2). We also discuss threats and possible alternatives (e.g., see [29]). Empirical studies show that sampling strategies can influence the number of detected faults or the precision of performance prediction [23,27].…”

Section: Related Workmentioning

confidence: 99%

Using machine learning to infer constraints for product lines

Temple

Galindo

Acher

et al. 2016

Proceedings of the 20th International Systems and Software Product Line Conference

View full text Add to dashboard Cite

Variability intensive systems may include several thousand features allowing for an enormous number of possible configurations, including wrong ones (e.g. the derived product does not compile). For years, engineers have been using constraints to a priori restrict the space of possible configurations, i.e. to exclude configurations that would violate these constraints. The challenge is to find the set of constraints that would be both precise (allow all correct configurations) and complete (never allow a wrong configuration with respect to some oracle). In this paper, we propose the use of a machine learning approach to infer such product-line constraints from an oracle that is able to assess whether a given product is correct. We propose to randomly generate products from the product line, keeping for each of them its resolution model. Then we classify these products accord-ing to the oracle, and use their resolution models to infer crosstree constraints over the product-line. We validate our approach on a product-line video generator, using a simple computer vision algorithm as an oracle. We show that an interesting set of cross-tree constraint can be generated, with reasonable precision and recall.

show abstract

“…In order to alleviate the inconveniences of this complex and costly task the scientific community has performed a constant effort [6]. Alshahrani and Peyravi presented a theoretical model to design and evaluate communication networks in data-centers [7].…”

Section: Listing 4 Data-center Configuration In Simcan 4 Evaluationmentioning

confidence: 99%

MAGICIAN: Model-based design for optimizing the configuration of data-centers

Cañizares

Núñez

Lara

2017

International Conferences on Software Engineering and Knowledge Engineering

View full text Add to dashboard Cite

Designing data-centers that provide an acceptable costperformance ratio is challenging. Generally, a wide spectrum of components must be previously analyzed, such as the kind of applications to be executed in the data-center, computing/storage requirements and the network topology, among others. Since each one of these components has a direct impact on the overall system performance, the design process is complex and difficult, which usually requires the intervention of an expert.We propose a model-based approach to design datacenters. For this purpose, we have created a meta-model that describes the structure of data-center models. Then, a set of expert rules can be used to detect sub-optimal configurations, and (in some cases) correct the design. Datacenter models can be simulated, to assess their performance and scalability, for which we use a code generator into the SIMCAN tool. We have implemented our approach as an Eclipse plugin, and illustrate the usefulness of some expert rules by showing the efficiency and scalability gains of the optimized model with respect to the original one.

show abstract

Performance-influence models for highly configurable systems

Cited by 202 publications

References 31 publications

Performance‐influence models of multigrid methods: A case study on triangular grids

Performance‐influence models of multigrid methods: A case study on triangular grids

Using machine learning to infer constraints for product lines

MAGICIAN: Model-based design for optimizing the configuration of data-centers

Contact Info

Product

Resources

About