REscope: High-dimensional statistical circuit simulation towards full failure region coverage

Wu, Wei; Xu, Weihua; Krishnan, Rahul; Chen, Yen-Lung; He, Lei

doi:10.1109/dac.2014.6881409

Cited by 6 publications

(12 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The proposed methods (SmartERA : without sample selection, SmartERA2: with sample selection) were implemented in Python. We performed the three different methods (MC, REscope [6], SmartERA, SmartERA2) to compare their accuracy and performance. …”

Section: Resultsmentioning

confidence: 99%

“…However, it is not capable of fully replacing the simulator due to its accuracy. So, a marginal filtering approach is used to improve the accuracy of classification [3]- [6]. This method uses relaxed threshold bounds instead of a real failure criterion to capture more samples to minimize classification error.…”

Section: B Classificationmentioning

confidence: 99%

“…Meanwhile, it is not sufficient to use a simple and linear classifier due to the nonlinearity of the failure region [3]- [5]. So, Gaussian radial basis function kernel (GRBF) and neural networks methods are available for nonlinear classifiers [6], [12]. Fig.…”

Section: B Classificationmentioning

confidence: 99%

“…Experimental result on the critical delay of 4-gates logic circuit with 48 dimensional process parameters, show that the proposed method can deliver 228x faster than the conventional MC method. Furthermore, the new method estimates 10x more accurately with only 1.2x additional simulation costs than the recently published approach [6].…”

Section: Introductionmentioning

confidence: 96%

“…Furthermore, the simulation cost from added samples from iterative process can be significant. Recently Wu et al [6] applied nonlinear SVM classifier to model nonlinear circuit behavior and multiple disjoint failure regions. Parameter pruning based on initial samples simulation was also applied for reducing the dimension of parameters so samples can only focus on sensitive parameters.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Rare event diagnosis by iterative failure region locating and elite learning sample selection

Shin

Tan

Shi

et al. 2015

2015 16th Latin-American Test Symposium (LATS)

View full text Add to dashboard Cite

Accurately estimating the failure region of rare events for nanoscale analog circuit blocks under process variations is a challenging task. In this paper, we propose a new statistical rare event analysis method. The new method is based on the iterative failure region locating scheme to reduce the sample counts while still maintains estimation accuracy. We derive the complete formulation for failure probability calculation in each iterative step. In addition, the new approach applies an elite learning sampling scheme, which considers both effectiveness of samples and well-coverage for process parameter search space, to further reduce number of samples after the gradual failure region locating at each iteration. As a result, the sample counts for simulation can be significantly reduced while sufficient representative samples are still kept for accurate failure probability analysis. Experiments were performed on the rise and fall time balancing failure analysis with Vth-NMOS, Vth-PMOS process parameters and the 4-gates logic circuit with 48 process parameters. The proposed method can deliver 228x faster than the conventional MC method with only 0.02-0.2% estimation errors. Furthermore, the new method estimates 20x more accurately with only 1.2x additional simulation costs than a recently published approach.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: B Classificationmentioning

confidence: 99%

Section: B Classificationmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 96%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Rare event diagnosis by iterative failure region locating and elite learning sample selection

Shin

Tan

Shi

et al. 2015

2015 16th Latin-American Test Symposium (LATS)

View full text Add to dashboard Cite

show abstract

Accurate Multi-segment Probability Density Estimation Through Moment Matching

Krishnan

Wu²,

Bodapati³

et al. 2024

IEEE Trans. Comput.-Aided Des. Integr. Circuits Syst.

Self Cite

View full text Add to dashboard Cite

In this manuscript, we present two novel algorithms for statistical circuit performance modeling.The first algorithm is a maximum entropy moment matching method (MAX-ENT) to model the overall distribution of the circuit performance considering process variations. A prior paper about this work was submitted by the same authors for publication in the ISQED-2013 proceedings, and cited in this manuscript with reference number [26]. This MAXENT algorithm is presented in Section III with more detailed derivation.In addition, we utilize the MAXENT approach, but apply it to a piecewise distribution model (PDM) framework, leading to a novel extension of the generic framework to model rare event failures in circuit modeling. We demonstrate why the prior work is not applicable to rare event circuit modeling and also provide a novel solution that showcases how the prior work can be abstracted for application to rare event circuit modeling. Furthermore, we provide several new experiments that demonstrate the advantage of using the current submissions framework over the prior papers framework. The PDM algorithm and experiment results are presented in Section IV and Section V respectively.Below is a brief summary of what is new to the manuscript and what was already proposed in the prior paper.Section I: Summary of basic statistical modeling algorithms and a brief overview of their strengths and weaknesses, including [26] Section II: A brief background on statistical modeling for circuits Section III: Detailed summary of the work proposed in [26], including derivations and experiment results (3 pages) Section IV: Completely new material (5 pages) that shows a new circuit modeling technique that extends the work in [26] by proposing new algorithms and models Section V: Completely new material (3 pages) that demonstrates the advantages of the current submissions algorithm over other algorithms including [26] The common denominator between [26] and the current submission is the fact that they utilize the same moment matching technique to obtain the probability distribution. However, although the underlying engine is the same, we estimate that at least 50% of the material is completely new because of the significant amount of new algorithms, models, and experimental results. 1 0278-0070 (c)Abstract-The impact of process variations continues to grow as transistor feature size shrinks. Such variations in transistor parameters lead to variations and unpredictability in circuit output and may ultimately cause them to violate specifications leading to circuit failure. In fact, timely failures in critical circuits may lead to catastrophic failures in the entire chip. As such, statistical modeling of circuit behavior is becoming increasingly important. However, existing statistical circuit simulation approaches fail to accurately and efficiently analyze the high sigma behavior of probabilistic circuit output. To this end, we propose PDM (Piecewise Distribution Model) -a piecewise distribution modeling approach via moment matching using ma...

show abstract

Hyperspherical Clustering and Sampling for Rare Event Analysis with Multiple Failure Region Coverage

Bodapati

2016

Proceedings of the 2016 on International Symposium on Physical Design

View full text Add to dashboard Cite

REscope: High-dimensional statistical circuit simulation towards full failure region coverage

Cited by 6 publications

References 24 publications

Rare event diagnosis by iterative failure region locating and elite learning sample selection

Rare event diagnosis by iterative failure region locating and elite learning sample selection

Accurate Multi-segment Probability Density Estimation Through Moment Matching

Hyperspherical Clustering and Sampling for Rare Event Analysis with Multiple Failure Region Coverage

Contact Info

Product

Resources

About