Die-to-Die Parasitic Extraction Targeting Face-to-Face Bonded 3D ICs
Author(s) -
Taigon Song,
Sung Kyu Lim
Publication year - 2015
Publication title -
journal of information and communication convergence engineering
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.28
H-Index - 6
eISSN - 2234-8883
pISSN - 2234-8255
DOI - 10.6109/jicce.2015.13.3.172
Subject(s) - face (sociological concept) , die (integrated circuit) , computer science , extraction (chemistry) , artificial intelligence , philosophy , chemistry , operating system , linguistics , chromatography
Face-to-face (F2F) bonding in three-dimensional integrated circuits (3D ICs), compared with other bonding styles, is closer to commercialization because of its benefits in terms of density, yield, and cost. However, despite the benefits that F2F bonding expect to provide, it’s physical nature has not been studied thoroughly. In this study, we, for the first time, extract cross-die (inter-die) parasitic elements from F2F bonds on the full-chip scale and compare them with the intra-die elements. This allows us to demonstrate the significant impact of field sharing across dies in F2F bonding on full-chip noise and critical path delay values. The baseline method used is the die-by-die method, where the parasitic elements of individual dies are extracted separately and the cross-die parasitic elements are ignored. Compared with this inaccurate method, which was the only method available until now, our first-of-its-kind holistic method corrects the delay error by 25.48% and the noise error by 175%.
Accelerating Research
Robert Robinson Avenue,
Oxford Science Park, Oxford
OX4 4GP, United Kingdom
Address
John Eccles HouseRobert Robinson Avenue,
Oxford Science Park, Oxford
OX4 4GP, United Kingdom