# US 10,762,166 B1 · Adaptive accelerated yield analysis > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 10,762,166 B1 - Document ID: US10762166B1 - Status: Granted - Kind: B1 - Application number: 16/434,334 - Filed: 2019-06-07 - Granted: 2020-09-01 - Earliest claimed: 2019-06-07 - Assignees: International Business Machines Corporation - Applicants: International Business Machines Corporation - Inventors: David Winston; Pravin Kamdar; Tong Li; Richard Daniel Kimmel - Web page: https://patentgrep.com/patent/US10762166B1 ## CPC - G06F 17/18: for evaluating statistical data {, e.g. average values, frequency distributions, probability functions, regression analysis (forecasting specially adapted for a specific administrative, business or logistic context G06Q10/04)} - G06N 20/10: using kernel methods, e.g. support vector machines [SVM] - G06F 30/367: Design verification, e.g. using simulation, simulation program with integrated circuit emphasis [SPICE], direct methods or relaxation methods - G06N 5/04: Inference or reasoning models - G06N 7/01: Probabilistic graphical models, e.g. probabilistic networks - G06N 20/00: Machine learning - G06F 30/27: using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model - G06F 2119/22: Yield analysis or yield optimisation - G06N 3/088: Non-supervised learning, e.g. competitive learning - G06N 5/01: Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound - G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N) - G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS ## Abstract Methods, systems, can computer program products for optimizing statistical method of computing output yields are provided. Aspects include determining a target criteria of a system for statistical analysis, based on the target criteria, determining a statistical analysis algorithm for the simulation, determining a block size for a plurality of statistical samples of the system for a parallelization of the statistical analysis algorithm, generating the plurality of statistical samples of the system, simulating the plurality of statistical samples of the system to determine one or more output yields, calculating a confidence interval for each of the one or more output yields, wherein the confidence interval comprises a lower bound, comparing the lower bound to a threshold standard deviation of a probability density function, and adjusting the block size for the plurality of statistical samples based on determining that the lower bound is less than the threshold standard deviation. ## Claims ### Claim 1 A computer-implemented method for optimizing statistical methods of computing output yields, the method comprising: determining a target criteria of a circuit design for statistical analysis; based at least in part on the target criteria, determining a statistical analysis algorithm for the statistical analysis; determining a block size for a plurality of statistical samples of the circuit design for a parallelization of the statistical analysis algorithm; generating the plurality of statistical samples of the circuit design; simulating the plurality of statistical samples of the circuit design to determine one or more output yields; calculating a confidence interval for each of the one or more output yields, wherein the confidence interval comprises a lower bound; comparing the lower bound to a threshold standard deviation of a probability density function; adjusting the block size for the plurality of statistical samples of the circuit design based on determining that the lower bound is less than the threshold standard deviation of the probability density function; generating a block of new statistical samples of the circuit design, wherein the block comprises the adjust block size; performing a simulation utilizing the block of new statistical samples of the circuit design; determining a result of the simulation; and manufacturing a circuit based on based on the result of the simulation. ### Claim 2 (depends on 1) The computer-implemented method of claim 1, further comprising: determining that the target criteria comprises a target failure probability; determining a failure rate for a first output yield of the one or more output yields, wherein the failure rate for the first output yield is determined based on comparing the first output yield to a failure threshold; based at least in part on the determining that the failure rate of the first output yield does not match the target failure probability, adjusting the failure threshold. ### Claim 3 (depends on 1) The computer-implemented method of claim 1, further comprising: determining that the target criteria comprises a target output yield; comparing a first output yield of the one or more output yields to the target output yield; based at least in part on determining that the first output yield is greater than the target output yield, removing the first output yield from the statistical analysis. ### Claim 4 (depends on 1) The computer-implemented method of claim 1, further comprising: determining a second statistical analysis algorithm based on determining that the lower bound is greater than the threshold standard deviation of the probability density function. ### Claim 5 (depends on 1) The computer-implemented method of claim 1, further comprising: analyzing, using a machine learning model, the plurality of statistical samples of the circuit design for the statistical analysis algorithm; predicting, by the machine learning model, one or more statistical samples of the plurality of samples of the circuit design that satisfy the target criteria; removing one or more target criteria from one or more target criteria of the circuit design. ### Claim 6 (depends on 5) The computer-implemented method of claim 5, further comprising: storing the one or more statistical samples in a memory; simulating the one or more statistical samples to determine output yields for the one or more sample configurations; and based on the output yields of the one or more of the target criteria, adjusting one or more parameters of the machine learning model. ### Claim 7 (depends on 1) The computer-implemented method of claim 1, further comprising: modifying the statistical analysis algorithm based on determining that the lower bound is greater than the threshold standard deviation of the probability density function. ### Claim 8 (depends on 1) The computer-implemented method of claim 1, wherein the confidence interval comprises a Wilson score interval. ### Claim 9 (depends on 1) The computer-implemented method of claim 1, wherein the probability density function comprises a normal distribution of historical output yields of historical simulations. ### Claim 10 (depends on 1) The computer-implemented method of claim 1, wherein the statistical analysis algorithm comprises a Monte Carlo method computational algorithm. ### Claim 11 (depends on 1) The computer-implemented method of claim 1, wherein the statistical analysis comprises a circuit simulation. ### Claim 12 A system for optimizing statistical methods of computing output yields, the system comprising: a processor communicatively coupled to a memory, the processor configured to: determine a target criteria of a circuit design for statistical analysis; based at least in part on the target criteria, determine a statistical analysis algorithm for the statistical analysis; determine a block size for a plurality of statistical samples of the circuit design for a parallelization of the statistical analysis algorithm; generate the plurality of statistical samples of the circuit design; simulate the plurality of statistical samples of the circuit design to determine one or more output yields; calculate a confidence interval for each of the one or more output yields, wherein the confidence interval comprises a lower bound; compare the lower bound to a threshold standard deviation of a probability density function; adjust the block size for the plurality of statistical samples of the circuit design based on determining that the lower bound is less than the threshold standard deviation of the probability density function; generate a block of new statistical samples of the circuit design, wherein the block comprises the adjust block size; perform a simulation utilizing the block of new statistical samples of the circuit design; and determine a result of the simulation; and manufacturing a circuit based on based on the result of the simulation. ### Claim 13 (depends on 12) The system of claim 12, wherein the processor is further configured to: determine that the target criteria comprises a target failure probability; determine a failure rate for a first output yield of the one or more output yields, wherein the failure rate for the first output yield is determined based on comparing the first output yield to a failure threshold; based at least in part on the determining that the failure rate of the first output yield does not match the target failure probability, adjust the failure threshold. ### Claim 14 (depends on 12) The system of claim 12, wherein the processor is further configured to: determine that the target criteria comprises a target output yield; compare a first output yield of the one or more output yields to the target output yield; based at least in part on determining that the first output yield is greater than the target output yield, remove the first output yield from the statistical analysis. ### Claim 15 (depends on 12) The system of claim 12, wherein the processor is further configured to: determine a second statistical analysis algorithm based on determining that the lower bound is greater than the threshold standard deviation of the probability density function. ### Claim 16 A computer program product for optimizing simulation output yields comprising a computer readable storage medium having program instructions embodied therewith, the program instructions executable by a processor to cause the processor to perform a method comprising: determining a target criteria of a circuit design for statistical analysis; based at least in part on the target criteria, determining a statistical analysis algorithm for the statistical analysis; determining a block size for a plurality of statistical samples of the circuit design for a parallelization of the statistical analysis algorithm; generating the plurality of statistical samples of the circuit design; simulating the plurality of statistical samples of the circuit design to determine one or more output yields; calculating a confidence interval for each of the one or more output yields, wherein the confidence interval comprises a lower bound; comparing the lower bound to a threshold standard deviation of a probability density function; adjusting the block size for the plurality of statistical samples of the circuit design based on determining that the lower bound is less than the threshold standard deviation of the probability density function; generating a block of new statistical samples of the circuit design, wherein the block comprises the adjust block size; performing a simulation utilizing the block of new statistical samples of the circuit design; determining a result of the simulation; and manufacturing a circuit based on based on the result of the simulation. ### Claim 17 (depends on 16) The computer program product of claim 16, further comprising: determining that the target criteria comprises a target failure probability; determining a failure rate for a first output yield of the one or more output yields, wherein the failure rate for the first output yield is determined based on comparing the first output yield to a failure threshold; based at least in part on the determining that the failure rate of the first output yield does not match the target failure probability, adjusting the failure threshold. ### Claim 18 (depends on 16) The computer program product of claim 16, further comprising: determining that the target criteria comprises a target output yield; comparing a first output yield of the one or more output yields to the target output yield; based at least in part on determining that the first output yield is greater than the target output yield, removing the first output yield from the statistical analysis. ### Claim 19 (depends on 16) The computer program product of claim 16, further comprising: determining a second simulation algorithm based on determining that the lower bound is greater than the threshold standard deviation of the probability density function. ### Claim 20 (depends on 16) The computer program product of claim 16, further comprising: analyzing, using a machine learning model, the plurality of statistical samples of the circuit design for the statistical analysis algorithm; predicting, by the machine learning model, one or more statistical samples of the plurality of statistical samples of the circuit design that satisfy the target criteria; removing the one or more statistical samples from one or more target criteria of the circuit design.