Synopsys, Inc.

Gung-Yu Pan

Chia-Chih Yen

Che-Hua Shih

Yi-An Chen

Chun Chan

Chia-Tung Chen

Shih-Ting Liu

Ssu-Hsien Li

Operations to recognize clock ports within a simulation circuit component and/or recognize a clock signal within simulation waveforms are described. One or more of the operations include generating a plurality of output values at an output port of a circuit simulation component by applying, during a simulation, a plurality of input values to a first input port of the circuit simulation component. The operations also include calculating a correlation vector based on bit sequences in the input values and bit sequences in the output values. The first input port is determined to be a clock port by applying a machine learning model to the correlation vector. One or more of the operations include determining a waveform file comprising signals from a simulation, determining a subset of the signals are bit-level signals, calculating toggle metrics for the subset of the signals, identifying a signal from the subset with a toggle metric satisfying a toggle threshold, calculating, by a processor, multiple duty cycles for the signal, and determining the signal is a clock signal based on the multiple duty cycles.

Data-driven clock port and clock signal recognition

A method of estimating a toggle count of a circuit, includes, in part, simulating the circuit to generate training data and an associated training toggle count of the internal nodes of the circuit in response to a test bench, training a neural network system to generate an estimate of the training toggle count in accordance with the training data and the associated training toggle count, simulating the circuit to generate simulation data in response to a first set of input values applied to the circuit, and invoking the trained neural network system to estimate a number of toggles of the internal nodes of the circuit from the simulation data. The training data may include, in part, values of input signals applied to the circuit and values of registers disposed in the circuit for a multitude of time stamps. The neural network system may include, in part, at least three layers.

Learning-based toggle estimation

A method for performing glitch power analysis of a circuit, comprising receiving no-timing waveform simulation data for the circuit, the waveform simulation data including a first signal, and identifying a delayed stimulus injection point (DSIP) for the first signal. The method further comprises determining a total delay for the first signal and performing waveform replay simulation including injecting the first signal at the DSIP at a time based on the total delay for the first signal.

Glitch analysis and glitch power estimation system

Growing community of inventors

Hsinchu, Taiwan

Gung-Yu Pan