An adaptive prioritized
            <i>ε</i>
            -preferred evolutionary algorithm for approximate BDD optimization

Shirinzadeh, Saeideh; Soeken, Mathias; Große, Daniel; Drechsler, Rolf

doi:10.1145/3071178.3071281

Cited by 12 publications

(10 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Our approach has produced more compact A-ROBDDs than the results reported in other earlier works on BDD approximation such as Shirinzadeh et al [2], [3], and Soeken et al [5], we've limited ourselves to presenting the comparison only with the state-of-the-art work [7] due to lack of space.…”

Section: Experiments and Discussionmentioning

confidence: 96%

“…Shirinzadeh et al have employed a genetic algorithm for synthesizing ROBDDs that represent an approximate Boolean function [2]. They use different approaches to optimize the multi-objective cost function that guides the search for an approximate ROBDD (A-ROBDD) within specified error constraints [2]- [4]. Soeken et al have made use of rounding operations for approximating Binary Decision Diagrams (BDDs) [5].…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

An Adaptive Metric for Node Substitution-Based BDD Minimization

Hassen

Khan

2021

2021 IEEE International Symposium on Circuits and Systems (ISCAS)

View full text Add to dashboard Cite

Modern computing workloads are becoming increasingly demanding and data-intensive. ASIC implementations designed for such computing workloads can consume valuable resources in terms of device size and system cost. Approximate Computing helps in preserving such resources in the case of applications that are error tolerant. We propose an Approximate Computing method that uses an adaptive heuristic to reduce the size of Reduced Ordered Binary Decision Diagrams through node substitutions. We establish the efficacy of our technique by showing improvement over a range of ISCAS-89 benchmarks.

show abstract

Section: Experiments and Discussionmentioning

confidence: 96%

Section: Related Workmentioning

confidence: 99%

An Adaptive Metric for Node Substitution-Based BDD Minimization

Hassen

Khan

2021

2021 IEEE International Symposium on Circuits and Systems (ISCAS)

View full text Add to dashboard Cite

show abstract

“…Unfortunately, the paper only experimentally shows the general applicability of these operations and does not present a full synthesis approach. In [19], an evolutionary algorithm working directly on the BDD representation of designs is proposed. It automatically adapts its preference for either the BDD size or the error metric value in order to ensure that no good solution is lost due to optimizing too much for a small error.…”

Section: B Approximate Logic Synthesismentioning

confidence: 99%

“…For BDD minimization, in general as well as in approximate settings, using genetic algorithms is a common approach, see, e.g., [10,19,29]. The optimization problem inherently has a multi-dimensional target space, minimizing both the BDD size and the error introduced through the approximation.…”

Section: Greedy Bucket-based Bdd Minimizationmentioning

confidence: 99%

BDD-Based Error Metric Analysis, Computation and Optimization

Keszöcze

2022

IEEE Access

View full text Add to dashboard Cite

Approximate Computing is a design paradigm that trades off computational accuracy for gains in non-functional aspects such as reduced area, increased computation speed, or power reduction. Computing the error of the approximated design is an essential step to determine its quality. The computation time for determining the error can become very large, effectively rendering the entire logic approximation procedure infeasible.In this work we extensively analyze various error metrics and approximation operations. We present methods to accelerate the computation of error metric computations by (a) exploiting structural information of the function obtained by applying the analyzed operations and (b) computing estimates of the metrics for multi-output Boolean functions represented as Binary Decision Diagrams (BDDs). We further present a novel greedy, bucket-based BDD minimization framework employing the newly proposed error metric computations to produce Pareto-optimal solutions with respect to BDD size and multiple error metrics. The applicability of the proposed minimization framework is demonstrated by an experimental evaluation. We can report considerable speedups while, at the same time, creating high-quality approximated BDDs. The presented framework is publicly available as open-source software on GitHub.

show abstract

“…The principles of approximate computing can be applied directly at the level of ROBDD circuit representation [35], i.e., before any circuit implementation is carried out. ROB-DDs are minimized under several objectives by performing both variable reordering and approximation while a predefined error constraint is not violated.…”

Section: Formal Error Analysis In the Fitness Functionmentioning

confidence: 99%

Evolutionary Algorithms in Approximate Computing: A Survey

Sekanina

2021

JICS

View full text Add to dashboard Cite

In recent years, many design automation methods have been developed to routinely create approximate implementations of circuits and programs that show excellent trade-offs between the quality of output and required resources. This paper deals with evolutionary approximation as one of the popular approximation methods. The paper provides the first survey of evolutionary algorithm (EA)-based approaches applied in the context of approximate computing. The survey reveals that EAs are primarily applied as multi-objective optimizers. We propose to divide these approaches into two main classes: (i) parameter optimization in which the EA optimizes a vector of system parameters, and (ii) synthesis and optimization in which EA is responsible for determining the architecture and parameters of the resulting system. The evolutionary approximation has been applied at all levels of design abstraction and in many different applications. The neural architecture search enabling the automated hardware-aware design of approximate deep neural networks was identified as a newly emerging topic in this area.

show abstract

An adaptive prioritized ε -preferred evolutionary algorithm for approximate BDD optimization

Cited by 12 publications

References 13 publications

An Adaptive Metric for Node Substitution-Based BDD Minimization

An Adaptive Metric for Node Substitution-Based BDD Minimization

BDD-Based Error Metric Analysis, Computation and Optimization

Evolutionary Algorithms in Approximate Computing: A Survey

Contact Info

Product

Resources

About