BEGIN:VCALENDAR VERSION:2.0 PRODID:IEEE vTools.Events//EN CALSCALE:GREGORIAN BEGIN:VTIMEZONE TZID:America/Toronto BEGIN:DAYLIGHT DTSTART:20220313T030000 TZOFFSETFROM:-0500 TZOFFSETTO:-0400 RRULE:FREQ=YEARLY;BYDAY=2SU;BYMONTH=3 TZNAME:EDT END:DAYLIGHT BEGIN:STANDARD DTSTART:20211107T010000 TZOFFSETFROM:-0400 TZOFFSETTO:-0500 RRULE:FREQ=YEARLY;BYDAY=1SU;BYMONTH=11 TZNAME:EST END:STANDARD END:VTIMEZONE BEGIN:VEVENT DTSTAMP:20220225T224411Z UID:E205F4EF-F21C-40EB-B739-9303F5547883 DTSTART;TZID=America/Toronto:20220225T160000 DTEND;TZID=America/Toronto:20220225T173000 DESCRIPTION:The design of advanced integrated circuits and microsystems fro m zero to terahertz frequencies calls for fast and accurate electromagneti cs-based modeling and simulation. The sheer complexity and high design cos t associated with the integrated circuits and microsystems prevent one fro m designing them based on hand calculation\, approximation\, intuition\, o r trial and error. The move towards higher frequencies and heterogeneous t echnologies stresses the need even more.\n\nHowever\, the analysis and des ign of integrated circuits (ICs) and microsystems impose many unique chall enges on electromagnetic analysis such as exponentially increased problem size and extremely multiscaled system spanning from nano- to centi-meter s cales. These challenges become new driving forces of the advancement of Co mputational Electromagnetics (CEM) in recent years\, since past techniques do not address them well.\n\nIn this talk\, recent advances in fast direc t solvers of O(N) (optimal) complexity will be presented\, including both direct PDE and IE solvers\, for addressing the ultra large problem size en countered in the IC design problems. In these solvers\, the underlying den se or sparse system matrix is directly inverted or factorized in O(N) comp lexity. To show how these solvers work\, a series of new accuracy controll ed fast matrix arithmetic will be elaborated including the representation of a dense matrix of O(N2) elements using O(N) parameters with controlled accuracy\, subsequent matrix-matrix multiplication\, matrix factorization\ , and inversion performed in O(N) complexity with directly controlled accu racy. The application of these fast algorithms to the design and analysis of industry product-level integrated circuits and systems will be presente d. Comparisons with direct and iterative solvers in the past will be made\ , which demonstrate the clear advantages of the new O(N) direct solvers.\n \nCo-sponsored by: Center for Computational Science and Engineering (CCSE) \, Faculty of Applied Science and Engineering\, University of Toronto\n\nS peaker(s): Dan Jiao\, \n\nVirtual: https://events.vtools.ieee.org/m/303190 LOCATION:Virtual: https://events.vtools.ieee.org/m/303190 ORGANIZER:costas.sarris@utoronto.ca SEQUENCE:6 SUMMARY:Fast Solvers for Electromagnetics-Based Analysis and Design of Inte grated Circuits and Systems URL;VALUE=URI:https://events.vtools.ieee.org/m/303190 X-ALT-DESC:Description: <br /><p>The design of advanced integrated circuits and microsystems from zero to terahertz frequencies calls for fast and ac curate electromagnetics-based modeling and simulation. The sheer complexit y and high design cost associated with the integrated circuits and microsy stems prevent one from designing them based on hand calculation\, approxim ation\, intuition\, or trial and error. The move towards higher frequencie s and heterogeneous technologies stresses the need even more.</p>\n<p>Howe ver\, the analysis and design of integrated circuits (ICs) and microsystem s impose many unique challenges on electromagnetic analysis such as expone ntially increased problem size and extremely multiscaled system spanning f rom nano- to centi-meter scales. These challenges become new driving force s of the advancement of Computational Electromagnetics (CEM) in recent yea rs\, since past techniques do not address them well.</p>\n<p>In this talk\ , recent advances in fast direct solvers of O(N) (optimal) complexity will be presented\, including both direct PDE and IE solvers\, for addressing the ultra large problem size encountered in the IC design problems. In the se solvers\, the underlying dense or sparse system matrix is directly inve rted or factorized in O(N) complexity. To show how these solvers work\, a series of new accuracy controlled fast matrix arithmetic will be elaborate d including the representation of a dense matrix of O(N2) elements using O (N) parameters with controlled accuracy\, subsequent matrix-matrix multipl ication\, matrix factorization\, and inversion performed in O(N) complexit y with directly controlled accuracy. The application of these fast algorit hms to the design and analysis of industry product-level integrated circui ts and systems will be presented. Comparisons with direct and iterative so lvers in the past will be made\, which demonstrate the clear advantages of the new O(N) direct solvers.</p> END:VEVENT END:VCALENDAR