In this talk, we will first discuss the need for computational electromagneucs (CEM) in hiah-frequcncy IC design. We will then review thc recent advancements in CEM, present chip design challenges to CEM analysis, and analyze the shortcomings of current CEM methodology. We will then proceed to emphasize the importance of chipdesigndriven CEM devclopment, and finally present some promising full-wave electromagnetic-based solutions that can achieve full-chip analysis and design. The Need for CEM

more »

... n Hieh-Frewencv IC Desim Circuit theory, as the low frequency p m of Maxwell's theory, has guided VLSI design and analysis far more than three decades. This theory is adequate as long as the physical dimension of the interconnects remains tiny compared to the wavelength of interest. Evolving into thc third decade, the clock frequency of microprocessors enters the gigahertz regime and heads towards ?OGHr. Since it is necessary to analyze thc chip response to hamonics 5 times the clock frequency and full wave effects are starting to be observable at about 1/100th of a wavelength, it is expected that interconnects longer than 3 0 p would have to be analyzed with certain full wave effects incorporated. As a result, computational electromagnetics, the science of solving Maxwell's equations at both low and high frequencies, has begun to draw anentian from on-chip design community.