International Business Machines Corporation

Globalfoundries Inc.

Emmanuel Yashchin

Nazmul Habib

Pascal A Nsame

Ernest Y Wu

John Greg Massey

Carole D Graas

Deborah M Massey

Yada Zhu

Peilin Song

Keith Aelwyn Jenkins

Barry P Linder

Franco Stellari

K Paul Muller

Daniel J Kearney

Alan Weger

James H Stathis

Raphael Peter Robertazzi

Emily A Ray

Kevin Geoffrey Stawiasz

Dimitri Kanevsky

Arun Kwangil Iyengar

Eric Vance Kline

Nianjun Zhou

Anthony Cappa Spielberg

Robert J Baseman

Jeffrey George Komatsu

Dhavalkumar C Patel

Paul Anthony Zulpa

Anuradha Bhamidipaty

Jingrui He

Andrew H Vogel

Benjamin J Steele, Jr

Shrey Shrivastava

Jonathan R M Hosking

Brian M Trapp

John M Wargo

Rejean P Levesque

Christina Landers

Asya Takken

Donnie Haye

Michael S Gordon

Ping-Chuan Wang

Arun Hampapur

Kenneth Parker Rodbell

Kai D Feng

Malcolm Scott Ware

Sara H Basson

Wei T Huang

William J Rooney

Peter Bergersen Yocom

Jayant R Kalagnanam

Zhijian Yang

Guillermo Jesus Silva

John Frederick Morar

Daniel A Oblinger

Sophia Krasikov

Scott J Hadderman

Ildar Khabibrakhmanov

Ilya Gluhovsky

David L Jensen

Herbert M Lee

Senthilnathan Velayudham

Alan Jerome Hoffman

Dimitri Kanevski

Vaughn Robert Evans

Sarah H Basson

Mary Yvonne Lanzerotti

Andrew J Dubois, Jr

Stephen Restivo

Christopher D Ross

Mary Wisniewski

Methods and systems to provide a form of probabilistic labeling to associate an outage with a disturbance, which could itself be either known based on the available data or unknown. In the latter case, labeling is especially challenging, as it necessitates the discovery of the disturbance. One approach incorporates a statistical change-point analysis to time-series events that correspond to service tickets in the relevant geographic sub-regions. The method is calibrated to separate the regular periods from the environmental disturbance periods, under the assumption that disturbances significantly increase the rate of loss-causing events. To obtain the probability that a given loss-causing event is related to an environmental disturbance, the method leverages the difference between the rate of events expected in the absence of any disturbances (baseline) and the rate of actually observed events. In the analysis, the local disturbances are identified and estimators of their duration and magnitude are provided.

Associating disturbance events to accidents or tickets

Disclosed embodiments provide techniques for estimating imputation algorithm performance. Multiple imputer algorithms are selected, and an evaluation of how well each of the imputer algorithms can estimate the missing data is performed. Disclosed embodiments obtain an imputer candidate dataset (ICD). The imputer candidate dataset is compared to the incomplete data range, and a similarity metric is determined between the data range and the ICD. When the similarity metric exceeds a predetermined threshold, an imputer evaluation dataset (IED) is created from the ICD by removing one or more data points from the ICD. Each imputer algorithm is evaluated by applying the IED to it, and computing an imputer evaluation metric based on its performance. The multiple imputer algorithms are ranked based on the imputer evaluation metric. The best ranked imputer algorithm can then be selected for use on the incomplete data range within the measurement dataset.

Data gap mitigation

Methods and systems of detecting chip degradation are described. A processor may execute a test on a device at a first time, where the test includes executable instructions for the device to execute a task under specific conditions relating to a performance attribute. The processor may receive performance data indicating a set of outcomes from the task executed by the device during the test. The processor may determine a first value of a parameter of the performance attribute based on the identified subset. The processor may compare the first value with a second value of the parameter of the performance attribute. The second value is based on an execution of the test on the device at a second time. The processor may determine a degradation status of the device based on the comparison of the first value with the second value.

Detection of performance degradation in integrated circuits

A per-chip equivalent oxide thickness (EOT) circuit sensor resides in an integrated circuit. The per-chip EOT circuit sensor determines electrical characteristics of the integrated circuit. The measured electrical characteristics include leakage current. The determined electrical characteristics are used to determine physical attributes of the integrated circuit. The physical attributes, including EOT, are used in a reliability model to predict per-chip failure rate.

Optimization of integrated circuit reliability

A computer-implemented method includes receiving a plurality of values of a data stream, where the plurality of values are generated by a monitored device and received in real time. One or more statistical moments of the data stream are updated, by a computer processing device, based on each value of the plurality of values. Each value of the plurality of values is discarded, after updating the one or more statistical moments of the data stream based on that value. A threshold is set for the data stream based on the one or more statistical moments. It is detected that the threshold has been passed by the data stream. A remedial action is performed on the monitored device, responsive to the threshold being passed, where the remedial action is associated with the threshold.

Real-time data analytics for streaming data

A method and system are provided for chip testing. The method includes selectively deploying a chip for future use or discarding the chip to prevent the future use, responsive to a stress history of the chip determined using a non-destructive test procedure. The test procedure includes ordering each of a plurality of functional patterns by a respective minimum operating period corresponding thereto. The test procedure further includes ranking each of the plurality of patterns based on at least one preceding available pattern to provide a plurality of pattern ranks. The test procedure also includes calculating a sum by summing the plurality of pattern ranks. The sum calculated during an initial performance of the test procedure is designated as a baseline, and the sum calculated during a subsequent performance of the test procedure is compared to a threshold derived from the baseline to determine the stress history of the chip.

Non-destructive analysis to determine use history of processor

A method and system are provided for chip testing. The method includes ascertaining a baseline for a functioning chip with no stress history by performing a non-destructive test procedure on the functioning chip. The method further includes repeating the test procedure on a chip under test using a threshold derived from the baseline as a reference point to determine a stress history of the chip under test. The test procedure includes ordering each of a plurality of functional patterns by a respective minimum operating period corresponding thereto, ranking each pattern based on at least one preceding available pattern to provide a plurality of pattern ranks, and calculating a sum by summing the pattern ranks. The sum calculated by the ascertaining step is designated as the baseline, and the sum calculated by the repeating step is compared to the threshold to determine the stress history of the chip under test.

An on-chip true noise generator including an embedded noise source with a low-voltage, high-noise zener diode(s), and an in-situ close-loop zener diode power control circuit. The present invention proposes the use of heavily doped polysilicon and silicon p-n diode(s) structures to minimize the breakdown voltage, increasing noise level and improving reliability. The present invention also proposes an in-situ close-loop zener diode control circuit to safe-guard the zener diode from catastrophic burn-out.

On-chip randomness generation

A system, method and computer program product for determining whether a material meets an alpha particle emissivity specification that includes measuring a background alpha particle emissivity for the counter and measuring a combined alpha particle emissivity from the counter containing a sample of the material. The combined alpha particle emissivity includes the background alpha particle emissivity in combination with a sample alpha particle emissivity. The decision statistic is computed based on the observed data and compared to a threshold value. When the decision statistic is less than the threshold value, the material meets the alpha particle emissivity specification. The testing times are computed based on pre-specified criteria so as to meet the needs of both Producer and Consumer.

Method of consumer/producer raw material selection

Growing community of inventors

Yorktown Heights, NY, United States of America

Emmanuel Yashchin