Cadence Design Systems, Inc.

Greg Horlick

Randall Scott Lawson

Ken Wadland

Richard Woodward

Sean Bergan

Walter M Katz

Paul S Musto

Donald Keith Morgan

Joe Smedley

An approach is provided for selectively optimizing a circuit design to be physical implemented. The approach includes generating a circuit routing solution in accordance with a plurality of constraints for parametric resources of the circuit design, with the constraints being defined respectively by a plurality of corresponding constraint instances. Each constraint instance variably indicates an effective constraining limit and degree of consumption for at least one of the parametric resources. At least one of the constraints is selectively adjusted by a predetermined over-constraining amount, and the circuit routing solution is preliminarily modified by applying at least one routing action selected responsive to the constraint adjustment. An automatic evaluation is then made of the potential impact upon constraint compliance. The evaluation includes generating a relative cost measure for the preliminary modification of the circuit routing solution, based at least partially upon each of the constraint instances. Thereafter, the preliminary modification of said circuit routing solution is either discarded or accepted responsive to the evaluation. A number of these operations may be repeated in recursive manner, if appropriate for the given application.

Method and system for optimized circuit autorouting

Automatic bundle filtering is provided to selectively configure a circuit design having a plurality of component terminals for physical implementation. A placement of components is established for a layout of the circuit design, and a plurality of connections to be routed between predetermined terminals of the components are defined for the layout, with a certain plurality of them selectively grouped into at least one candidate bundle. At least one filter is applied to the connections of each candidate bundle for responsive segregation according to a preselected connection discriminant into one or more updated bundle candidates. Each updated bundle candidate is preferably evaluated in accordance with at least one preselected bundling criteria to identify acceptable bundle candidates therefrom. Bundles corresponding to the acceptable bundle candidates are then generated.

Method and system for adaptive bundling of connections in user-guided autorouting

An approach is provided for selectively optimizing a circuit design, including generating a circuit routing solution according to a plurality of constraints for parametric resources of the circuit design, with the constraints being defined respectively by a plurality of corresponding constraint instances. Each constraint instance variably indicates an effective constraining limit and degree of consumption for at least one of the parametric resources. At least one of the constraints is selectively adjusted by a predetermined over-constraining amount, and the circuit routing solution is preliminarily modified by applying at least one routing action selected responsive to the constraint adjustment. The potential impact upon constraint compliance is evaluated, including generating a relative cost measure for the preliminary modification of the circuit routing solution, based at least partially upon each of the constraint instances. Thereafter, the preliminary modification of the circuit routing solution is either discarded or accepted responsive to the evaluation.

A mechanism is provided for the user to define a circuit design intent or strategy in the form of data that is stored with the design database. An autorouter then uses this guidance from the user to create a plan for routing the design. The user can then modify their guidance to the router until the results for the plan are acceptable. Using the planned flow, the autorouter can complete the design, creating detailed paths including etch segments and vias. Allowing such interaction with an autorouter significantly reduces the routing time and hence time-to-market.

Growing community of inventors

Madison, AL, United States of America

Greg Horlick