Abstract: This application discloses a computing system to identify suspected defects in a manufactured integrated circuit, which correspond to electrical failures detected by a test applied to the manufactured integrated circuit. The computing system can utilize the suspected defects in the manufactured integrated circuit to cluster features in a physical layout design describing the manufactured integrated circuit. Each cluster of the features corresponds to a candidate for a physical root cause of the suspected defects in the manufactured integrated circuit. The computing system can detect a physical root cause of the electrical failures in the manufactured integrated circuit based on the clusters of the features. A physical failure analysis process includes an inspection of the manufactured integrated circuit to confirm the physical root cause of the electrical failures in the manufactured integrated circuit corresponds to a systemic manufacturing fault in the manufactured integrated circuit.

Abstract: Various aspects of the disclosed technology relate to machine learning-based chain diagnosis. Faults are injected into scan chains in a circuit design. Simulations are performed on the fault-injected circuit design to determine test response patterns in response to the test patterns which are captured by the scan chains. Observed failing bit patterns are determined by comparing the unloaded test response patterns with corresponding good-machine test response patterns. Bit-reduction is performed on the observed failing bit patterns to construct training samples. Using the training samples, machine-learning models for faulty scan cell identification are trained. The bit reduction comprises pattern-based bit compression for good scan chains or cycle-based bit compression for the good scan chains. The bit reduction may further comprise bit-filtering. The bit-filtering may comprises keeping only sensitive bits on faulty scan chains for the training samples construction.

Abstract: Various aspects of the disclosed technology relate to machine learning-based chain diagnosis. Faults are injected into scan chains in a circuit design. Simulations are performed on the fault-injected circuit design to determine observed failing bit patterns. Bit-reduction is performed on the observed failing bit patterns to construct first training samples. Using the first training samples, first-level machine-learning models are trained. Affine scan cell groups are identified. Second training samples are prepared for each of the affine scan cell groups by performing bit-filtering on a subset of the observed failing bit patterns associated with the faults being injected at scan cells in the each of the affine scan cell groups. Using the second training samples, second-level machine-learning models are trained. The first-level and second-level machine learning models can be applied in a multi-stage machine learning-based chain diagnosis process.

Abstract: A computing system may include a model training engine configured to train a supervised learning model with a training set comprising training probability distributions computed for training dies through a local phase of a volume diagnosis procedure. The computing system may also include a volume diagnosis adjustment engine configured to access a diagnosis report for a given circuit die that has failed scan testing and compute, through the local phase of the volume diagnosis procedure, a probability distribution for the given circuit die from the diagnosis report. The volume diagnosis adjustment engine may also adjust the probability distribution into an adjusted probability distribution using the supervised learning model and provide the adjusted probability distribution for the given circuit die as an input to a global phase of the volume diagnosis procedure to determine a global root cause distribution for multiple circuit dies that have failed the scan testing.

Abstract: Logic diagnosis is performed on failing reports of defective integrated circuits to derive a diagnosis report for each of the failing reports which comprise information of suspects. The suspects comprise cell internal suspects and interconnect suspects. A probability distribution of root causes for causing the defective integrated circuits is determined to maximize a likelihood of observing the diagnosis reports based on a probability for each of the suspects given each of the root causes and a probability for each of the diagnosis reports given each of the suspects. The probability for each of the diagnosis reports given each of the cell internal suspects is weighted higher than the probability for each of the diagnosis reports given each of the interconnect suspects.

Abstract: One or more machine-learning models are trained and employed to predict test coverage and test data volume. Input features for the one or more machine-learning models comprise the test configuration features and the design complexity features. The training data are prepared by performing test pattern generation and circuit design analysis. The design complexity features may comprise testability, X-profiling, clock domains, power domains, design-rule-checking warnings, or any combination thereof.

Abstract: Various aspects of the disclosed technology relate to machine learning-based chain diagnosis. Faults are injected into scan chains in a circuit design. Simulations are performed on the fault-injected circuit design to determine observed failing bit patterns. Bit-reduction is performed on the observed failing bit patterns to construct first training samples. Using the first training samples, first-level machine-learning models are trained. Affine scan cell groups are identified. Second training samples are prepared for each of the affine scan cell groups by performing bit-filtering on a subset of the observed failing bit patterns associated with the faults being injected at scan cells in the each of the affine scan cell groups. Using the second training samples, second-level machine-learning models are trained. The first-level and second-level machine learning models can be applied in a multi-stage machine learning-based chain diagnosis process.

Abstract: Various aspects of the disclosed technology relate to machine learning-based chain diagnosis. Faults are injected into scan chains in a circuit design. Simulations are performed on the fault-injected circuit design to determine test response patterns in response to the test patterns which are captured by the scan chains. Observed failing bit patterns are determined by comparing the unloaded test response patterns with corresponding good-machine test response patterns. Bit-reduction is performed on the observed failing bit patterns to construct training samples. Using the training samples, machine-learning models for faulty scan cell identification are trained. The bit reduction comprises pattern-based bit compression for good scan chains or cycle-based bit compression for the good scan chains. The bit reduction may further comprise bit-filtering. The bit-filtering may comprises keeping only sensitive bits on faulty scan chains for the training samples construction.