A parallel analog circuit automatic optimization method based on genetic algorithm and machine learning comprises global optimization based on genetic algorithm and local optimization based on machine learning, with the global optimization and the local optimization performed alternately. The global optimization based on genetic algorithm utilizes parallel SPICE simulations to improve the optimization efficiency while guaranteeing the optimization accuracy, combined with parallel computing. The local optimization based on machine learning establishes a machine learning model near the global optimal point obtained by the global optimization, and uses the machine learning model to replace the SPICE simulator, thus reducing the time costs brought by a large number of simulations.

Systems and methods related to learning-based analyzers (both supervised and unsupervised) for mitigating latch-up in integrated circuits are provided. An example method includes obtaining latch-up data concerning at least one integrated circuit configured to operate under a range of temperature conditions, where the at least one integrated circuit comprises a core portion including at least a plurality of devices each having one or more structural features formed using a lithographic process, and an input/output portion. The method further includes training the learning-based system based on training data derived from the latch-up data and a first layout rule concerning a first spacing between the core portion and the input/output portion. The method further includes using the learning-based system generating a second layout rule concerning the first spacing between the core portion and the input/output portion, where the second layout rule is different from the first layout rule.

Systems and methods for computer-assisted design of an inductor are described. Target specifications for an inductor are received. An inductor design is generated segment-by-segment using a reinforcement learning agent to generate segment parameters for each added segment. The reinforcement learning agent implements a policy that is learned using a reward computed based on performance of the generated inductor design relative to the target specifications. The generated inductor design is outputted as a candidate inductor design after determining that the generated inductor design satisfies a predefined performance threshold.

The present disclosure relates to routing superconducting wires in superconducting circuits and in particular to efficiently routing superconducting wires that meet inductance requirements. The superconducting wire routing technique involves modeling the target location not only as a physical location, but as a physical location (e.g., x, y, and z dimensions) combined with inductance (e.g., a target inductance range). One or more other constraints may also be included in the modeling, such as a number of wires that would need to be moved/lifted, a number of circuit-vias allowing passage through layers of the circuit, an amount of cross-coupling with other inductors, and a number of wire segments.

Methods for modifying power use of a semiconductor device include receiving, at one or more processors, an activity stream of a simulation of a semiconductor device, the activity stream comprising a stream of signals. Using the one or more processors, integrated circuit actions are recognized from the activity stream, each integrated circuit action representing an abstraction of work done by the semiconductor device. The processor(s) determine one or more values associated with the integrated circuit actions. A model of power use is generated for the semiconductor device, the model based at least in part on the recognized integrated circuit actions and the associated values. Based on an output of the model, power use of the semiconductor device is modified. Other methods and systems related to determining, modeling, and predicting power/energy use of semiconductor devices are also disclosed.

Methods for modifying power use of a semiconductor device include receiving, at one or more processors, an activity stream of a simulation of a semiconductor device, the activity stream comprising a stream of signals. Using the one or more processors, integrated circuit actions are recognized from the activity stream, each integrated circuit action representing an abstraction of work done by the semiconductor device. The processor(s) determine one or more values associated with the integrated circuit actions. A model of power use is generated for the semiconductor device, the model based at least in part on the recognized integrated circuit actions and the associated values. Based on an output of the model, power use of the semiconductor device is modified. Other methods and systems related to determining, modeling, and predicting power/energy use of semiconductor devices are also disclosed.

A technique for designing circuits including receiving a data object representing a circuit for a first process technology, the circuit including a first sub-circuit, the first sub-circuit including a first electrical component and a second electrical component arranged in a first topology; identifying the first sub-circuit in the data object by comparing the first topology to a stored topology, the stored topology associated with the first process technology; identifying a first set of physical parameter values associated with first electrical component and the second electrical component of the first sub-circuit; determining a set of performance parameter values for the first sub-circuit based on a first machine learning model of the first sub-circuit and the identified set of physical parameters; converting the identified first sub-circuit to a second sub-circuit for the second process technology based on the determined set of performance parameter values; and outputting the second sub-circuit.

A technique for designing circuits including receiving a data object representing a circuit for a first process technology, the circuit including a first sub-circuit, the first sub-circuit including a first electrical component and a second electrical component arranged in a first topology; identifying the first sub-circuit in the data object by comparing the first topology to a stored topology, the stored topology associated with the first process technology; identifying a first set of physical parameter values associated with first electrical component and the second electrical component of the first sub-circuit; determining a set of performance parameter values for the first sub-circuit based on a first machine learning model of the first sub-circuit and the identified set of physical parameters; converting the identified first sub-circuit to a second sub-circuit for the second process technology based on the determined set of performance parameter values; and outputting the second sub-circuit.

Example embodiments relate to a method for designing integrated circuit based on a computer program including at least one instruction performed by an integrated circuit design system including a processor, the method comprising mapping, by the processor, at least one netlist signal to at least one RTL signal, determining, by the processor, a target netlist signal among the at least one netlist signal, determining, by the processor, a target expression of the target netlist signal using the netlist signal mapped to the at least one RTL signal based on at least a part of the netlist circuit, simulating, by the processor, on the at least one RTL signal, calculating, by the processor, signal simulation predicting information for the target netlist signal using the target expression and generating, by the processor, a design model by designing an integrated circuit using the signal simulation predicting information.

Example embodiments relate to a method for designing integrated circuit based on a computer program including at least one instruction performed by an integrated circuit design system including a processor, the method comprising mapping, by the processor, at least one netlist signal to at least one RTL signal, determining, by the processor, a target netlist signal among the at least one netlist signal, determining, by the processor, a target expression of the target netlist signal using the netlist signal mapped to the at least one RTL signal based on at least a part of the netlist circuit, simulating, by the processor, on the at least one RTL signal, calculating, by the processor, signal simulation predicting information for the target netlist signal using the target expression and generating, by the processor, a design model by designing an integrated circuit using the signal simulation predicting information.