An efficient partitioning and placement based fault TSV detection in 3D-IC using deep learning approach

Basit Öğe Kaydı
dc.authorid, Dr.P.Sivakumar/0000-0003-1328-8093
dc.authoridCengiz, Korhan/0000-0001-6594-8861
dc.authoridCengiz, Korhan/0000-0001-6594-8861
dc.authoridRADHAKRISHNAN NAIR, RADEEP KRISHNA/0000-0003-3576-2032
dc.authorwosid, Dr.P.Sivakumar/AAC-5503-2019
dc.authorwosidCengiz, Korhan/ABD-5559-2020
dc.authorwosidCengiz, Korhan/HTN-8060-2023
dc.contributor.authorRadhakrishnan Nair, Radeep Krishna
dc.contributor.authorPothiraj, Sivakumar
dc.contributor.authorRadhakrishnan Nair, T. R.
dc.contributor.authorCengiz, Korhan
dc.date.accessioned2024-06-12T11:16:19Z
dc.date.available2024-06-12T11:16:19Z
dc.date.issued2021
dc.departmentTrakya Üniversitesien_US
dc.description.abstractOver topical eras, three dimensional Integrated Circuit (3D-IC) fabrications have become vital among the researchers and industrial people, owing to its wide range of amenities including smaller intersect lengths, advanced incorporation density, and enhanced performance. Still, fault Through Silicon Via (TSV) detection is a bottleneck, due to poor fabrication processes such as partitioning and placement. Besides, state of the art works have concentrated on redundant TSV allocation instead of detected fault TSV and hence, the area overhead and size of the circuit are increased. To resolve these shortcomings, this paper proposes an Efficient Partitioning and Placement based Fault TSV detection in 3D-IC. The proposed work comprises five processes: Quick cut oriented Partitioning, Multi-Objective based Placement, Deep learning based Fault TSV detection, Re-routing and Adaptive Time Division Multiple Access (TDMA) time slot. Initially, Quick Cut algorithm has been employed to partition the 3D-IC and it is easier for placement process. The placement is executed through Multi-Objective Brain Storm Optimization algorithm that selects the optimal place to position the cells in 3D-IC. The fault TSV in the 3D-IC is detected using the Adam Deep Neural Network algorithm. Further, Adam optimizer has been used to estimate weight for each input and it provides fast performance and better convergence rate compared to the traditional stochastic gradient algorithm. After obtaining the fault TSV, rerouting is performed to reroute the signals transmitted over the defected TSV to the nearby defect free TSV. The Adaptive TDMA algorithm has been used to provide time slot to TSV positioned in each partition. The proposed method has been implemented in MATLABR2017b tool. The results attained from the simulations are propitious in terms of the metrics such as Area, Wirelength, Delay, Run time and Temperature.en_US
dc.identifier.doi10.1007/s12652-021-02964-w
dc.identifier.issn1868-5137
dc.identifier.issn1868-5145
dc.identifier.scopus2-s2.0-85102257088en_US
dc.identifier.scopusqualityQ1en_US
dc.identifier.urihttps://doi.org/10.1007/s12652-021-02964-w
dc.identifier.urihttps://hdl.handle.net/20.500.14551/24277
dc.identifier.wosWOS:000625581200001en_US
dc.identifier.wosqualityQ2en_US
dc.indekslendigikaynakWeb of Scienceen_US
dc.indekslendigikaynakScopusen_US
dc.language.isoenen_US
dc.publisherSpringer Heidelbergen_US
dc.relation.ispartofJournal Of Ambient Intelligence And Humanized Computingen_US
dc.relation.publicationcategoryMakale - Uluslararası Hakemli Dergi - Kurum Öğretim Elemanıen_US
dc.rightsinfo:eu-repo/semantics/closedAccessen_US
dc.subjectThree Dimensional Integrated Circuiten_US
dc.subjectQuick Cuten_US
dc.subjectPlacementen_US
dc.subjectFault Through Silicon Viaen_US
dc.subjectRe-Routingen_US
dc.subjectAdaptive Time Sloten_US
dc.titleAn efficient partitioning and placement based fault TSV detection in 3D-IC using deep learning approachen_US
dc.typeArticleen_US

Dosyalar

Koleksiyon

WoS İndeksli Yayınlar Koleksiyonu
Scopus İndeksli Yayınlar Koleksiyonu