Scalable Modeling

Overview

Foundries typically provide PDKs with a basic set of passive devices that can response to user’s input of dimension parameters.

Pcells are written in SKILL (or alterative PyCell). Scalable models are traditionally derived based on huge amount of silicon data. EM solver based modeling has been used to deliver simulated models. Use Cadence built-in features in offering device parameter changes and netlisting support.

A scalable model is a series of empirical relations between the circuit model elements and geometric parameters.ScalableModel

PeakView scalable model package includes the following components:

A scalable model generation package with each step automated.

Customers only need to provide the profile as input and receive the scalable model netlist for a process node.

➢      Input: PeakView profile for the technology process

➢      Output: Scalable model netlist for the corresponding technology process

Data flow is shown below:

ScalableModel2

PeakView platform is a natural fit for scalable model development. It has the accuracy and speed of a 3D EM solver. PCircuits generates the layout for all kinds of passive devices with user controlled geometrical parameters. LEM can take the layout from PCircuit, GDS and Pcells. PeakView’s PBM is a perfect basis to build scalable model. The scalable model generation flow is highly automated in the flexible and powerful platform.

Benefits

  1. Writing PCircuit is easy
  2. PBM is customizable
  3. Model for new devices can be quickly developed
  4. Flow is highly automated

Comments are closed