A Spectral Approach to Scalable Vectorless Thermal Integrity Verification

Zhao, Zhiqiang; Feng, Zhuo

doi:10.23919/date48585.2020.9116438

Cited by 4 publications

(3 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To achieve the desired levels of chip reliability and functionality, compute-intensive full-chip thermal analysis and integrity verifications are indispensable, which typically involves estimating thermal profiles under a variety of workloads and power budgets. In this work, we introduce a data-driven vectorless power/thermal integrity verification framework: (1) given a collection of voltage/temperature measurements that can be potentially obtained from on-chip voltage/temperature sensors [32], [33], the proposed data-driven method will first construct a sparse power/thermal grid network leveraging the proposed graph topology learning approach; (2) next, vectorless power/thermal integrity verification framework will be exploited for estimating the worst-case voltage/temperature (gradient) distributions [53], [54].…”

Section: Data-driven Vectorless Integrity Verificationmentioning

confidence: 99%

SF-SGL: Solver-Free Spectral Graph Learning from Linear Measurements

Zhang,

Zhao,

Feng

2023

Preprint

View full text Add to dashboard Cite

This work introduces a highly-scalable spectral graph densification framework (SGL) for learning resistor networks with linear measurements, such as node voltages and currents. We show that the proposed graph learning approach is equivalent to solving the classical graphical Lasso problems with Laplacian-like precision matrices. We prove that given O(log N ) pairs of voltage and current measurements, it is possible to recover sparse N -node resistor networks that can well preserve the effective resistance distances on the original graph. In addition, the learned graphs also preserve the structural (spectral) properties of the original graph, which can potentially be leveraged in many circuit design and optimization tasks.To achieve more scalable performance, we also introduce a solver-free method (SF-SGL) that exploits multilevel spectral approximation of the graphs and allows for a scalable and flexible decomposition of the entire graph spectrum (to be learned) into multiple different eigenvalue clusters (frequency bands). Such a solver-free approach allows us to more efficiently identify the most spectrally-critical edges for reducing various ranges of spectral embedding distortions. Through extensive experiments for a variety of real-world test cases, we show that the proposed approach is highly scalable for learning sparse resistor networks without sacrificing solution quality. We also introduce a datadriven EDA algorithm for vectorless power/thermal integrity verifications to allow estimating worst-case voltage/temperature (gradient) distributions across the entire chip by leveraging a few voltage/temperature measurements.

show abstract

Section: Data-driven Vectorless Integrity Verificationmentioning

confidence: 99%

SF-SGL: Solver-Free Spectral Graph Learning from Linear Measurements

Zhang,

Zhao,

Feng

2023

Preprint

View full text Add to dashboard Cite

show abstract

“…Hence, the majority of the thermal analysis approaches rely on the vector-based simulations [38,60]. Recently, [115] proposed the first vectorless thermal verification algorithm, which can be easily scaled to very large scale thermal grid designs.…”

Section: A Spectral Approach To Scalable Vectorless Power Grid and Thmentioning

confidence: 99%

“…Motivated by the existing vectorless integrity verification problems [25,33,105,107,113,115], we propose the first general vectorless integrity verification framework which can be applied to both power and thermal grids to provide the scalable solutions for estimating the maximum voltage drop or the nearly-worst-case thermal profiles under various complex power density or workload uncertainties and constraints. It leverages a recent graph-theoretic algebraic multigrid (AMG) algorithmic framework [64] as well as a hierarchy of almost linear-sized sparsifiers.…”

Section: A Spectral Approach To Scalable Vectorless Power Grid and Thmentioning

confidence: 99%