Near-Optimal Thermal Monitoring Framework for Many-Core Systems-on-Chip

Ranieri, Juri; Vincenzi, Alessandro; Chebira, Amina; Atienza, David; Vetterli, Martin

doi:10.1109/tc.2015.2395423

Cited by 21 publications

(15 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The final estimate of the posterior distribution (2) can be obtained by starting from any initial guess for the model parameters, H 0 = {Γ 0 , Λ 0 } and B 0 , and iterating between update rules (7), (10), and (2) until convergence. We call the resulting algorithm full Dugh.…”

Section: Proposed Methods -Full Dughmentioning

confidence: 99%

“…Probabilistic graphical models for regression problems, where both the model parameters and the observation noise can have complex spatio-temporal correlation structure, are prevalent in neuroscience, neuroimaging and beyond [1]. Example domains where these models are being applied include Gaussian process inference [2], sparse signal recovery and multi-task learning [3][4][5][6], array signal processing [7], fMRI analysis [8], climate science [9], computer hardware monitoring [10], and brain-computer interfaces (e.g. [11,12]).…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Efficient Hierarchical Bayesian Inference for Spatio-temporal Regression Models in Neuroimaging

Hashemi¹,

Gao²,

Cai³

et al. 2021

Preprint

View full text Add to dashboard Cite

Several problems in neuroimaging and beyond require inference on the parameters of multi-task sparse hierarchical regression models. Examples include M/EEG inverse problems, neural encoding models for task-based fMRI analyses, and climate science. In these domains, both the model parameters to be inferred and the measurement noise may exhibit a complex spatio-temporal structure. Existing work either neglects the temporal structure or leads to computationally demanding inference schemes. Overcoming these limitations, we devise a novel flexible hierarchical Bayesian framework within which the spatio-temporal dynamics of model parameters and noise are modeled to have Kronecker product covariance structure. Inference in our framework is based on majorization-minimization optimization and has guaranteed convergence properties. Our highly efficient algorithms exploit the intrinsic Riemannian geometry of temporal autocovariance matrices. For stationary dynamics described by Toeplitz matrices, the theory of circulant embeddings is employed. We prove convex bounding properties and derive update rules of the resulting algorithms. On both synthetic and real neural data from M/EEG, we demonstrate that our methods lead to improved performance.

show abstract

Section: Proposed Methods -Full Dughmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Efficient Hierarchical Bayesian Inference for Spatio-temporal Regression Models in Neuroimaging

Hashemi¹,

Gao²,

Cai³

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…Monitoring temperature over spatial domains is an important task with practical importance for surveillance and automation purposes [19]. Areas of application include agriculture [1], climate change studies [39], thermal monitoring of CPUs [33], and environmental protection [37]. Other applications include monitoring of the thermal field caused by on-chip sensors for many-core systems [33,29], and cultivation of sensitive species [16].…”

Section: Introductionmentioning

confidence: 99%

“…Areas of application include agriculture [1], climate change studies [39], thermal monitoring of CPUs [33], and environmental protection [37]. Other applications include monitoring of the thermal field caused by on-chip sensors for many-core systems [33,29], and cultivation of sensitive species [16]. For these and similar applications, it is necessary to monitor the temperature with high resolution using a proper sampling device for possibly long term periods.…”

Section: Introductionmentioning

confidence: 99%

Compressed Sensing: From Theory to Praxis

Flinth¹,

Hashemi²,

Kutyniok³

2017

Compressive Sensing of Earth Observations

View full text Add to dashboard Cite

We propose a scheme utilizing ideas from infinite dimensional compressed sensing for thermal source localization. Using the soft recovery framework of one of the authors, we provide rigorous theoretical guarantees for the recovery performance. In particular, we extend the framework in order to also include noisy measurements. Further, we conduct numerical experiments, showing that our proposed method has strong performance, in a wide range of settings. These include scenarios with few sensors, off-grid source positioning and high noise levels, both in one and two dimensions.objective function based on the Unit-Norm-Tight-Frame (UNTF) concept. Although the paper emphasizes on empirical aspects, theoretical claims are also derived.Excluding [23,22], all of the previously mentioned works do not explicitly try to utilize the useful structures which exist for diffusion sources, such as sparsity and spatiotemporal correlation among measurement governed by a PDE constraint. In particular, none of them use the powerful framework of compressed sensing [2]. Therefore, Rostami et al. [34] proposed a compressed sensing method for reconstructing the diffusion field by incorporating PDE constraints in recovery part as a side information. This effort was extended to the 2D scenario in [15] by one of the authors of this article, together with co-authors. In a similar way but by utilizing an analysis formulation for the source localization problem, [17] consider the source localization inverse problem for other types of sources or governing equations, with PDE side information in a co-sparse framework.

show abstract

“…此外, Reda等人 [21] 通过启发式搜索技术优化传感器布置并结合线性加权方法来估计间接测量点温度, 从而实现提高温度监测分辨率的目的. 有学者基于框架理论和贪心多项式算法重构温度矩阵 [22,23] .…”

unclassified

The implementation of software and hardware for dynamic thermal management of electronic devices

Wang¹,

Pan²,

Wu³

et al. 2020

Sci. Sin.-Tech.

View full text Add to dashboard Cite

Near-Optimal Thermal Monitoring Framework for Many-Core Systems-on-Chip

Cited by 21 publications

References 28 publications

Efficient Hierarchical Bayesian Inference for Spatio-temporal Regression Models in Neuroimaging

Efficient Hierarchical Bayesian Inference for Spatio-temporal Regression Models in Neuroimaging

Compressed Sensing: From Theory to Praxis

The implementation of software and hardware for dynamic thermal management of electronic devices

Contact Info

Product

Resources

About