| The relentless pursuit of techniques to improve density, speed, and power consumption of digital integrated circuits has been systematically reducing the noise rejection ability of the integrated circuit technology while simultaneously increasing the amount of noise being generated. These two trends have reached a point where the natural noise rejection ability of the technology is no longer sufficient to ensure error-free operation of the systems that are being implemented with the technology. We now must take an active role in managing noise issues when designing new systems in order to guarantee that the noise levels do not exceed the noise rejection ability of the technology. In this dissertation we develop the key elements of a new post-route crosstalk noise management strategy.; To begin, we formulate the heart of our strategy as a nonlinear convex programming problem. Initially we base our formulation on Devgan's crosstalk noise estimate [1]. This estimate allows us to easily formulate a convex programming problem which accounts for the effects of coupling capacitance, victim net interconnect resistance, victim net source resistance, as well as aggressor net rise time on nets with arbitrary tree topologies. We then develop several efficient methods for solving variations of this programming problem.; Although Devgan's estimate is a substantial improvement over previous estimation techniques, it does not account for all crosstalk noise mechanisms. However, the increased complexity of more sophisticated estimates lead to a crosstalk management problem that is very hard to solve. We develop two new estimates, similar in form to Devgan's estimate, which are based on local approximations of Vittal's estimate [2]. Our estimates are substantially more accurate than Devgan's estimate while still allowing us to formulate the crosstalk management problem in a form that can be solved efficiently.; Finally, we present a new post-route routing resource manager which is based on a routing resource migration strategy. Our new approach moves uncommitted routing resources through the routing so that they can be concentrated around specific routing elements while minimizing the disruption to the rest of the routing. |
