8.6 Statistical Answers to Analog/Mixed Signal Design and Test Problems
Date: Wednesday 11 March 2015
Time: 17:00 - 18:30
Location / Room: Bayard
Chair:
Jacob Abraham, The University of Texas at Austin, US
Co-Chair:
Michel Renovell, LIRMM/CNRS, FR
The session will demonstrate applications of Bayesian model fusion, machine learning classifiers, feature selection, virtual probe, and Quasi Monte Carlo for solving challenging design and test problems for analog and mixed signal circuits.
| Time | Label | Presentation Title Authors |
|---|---|---|
| 17:00 | 8.6.1 | EFFICIENT BIT ERROR RATE ESTIMATION FOR HIGH-SPEED LINK BY BAYESIAN MODEL FUSION Speakers: Chenlei Fang1, Qicheng Huang1, Fan Yang1, Xuan Zeng1, Xin Li2 and Chenjie Gu3 1Fudan University, CN; 2Carnegie Mellon University and Fudan University, US; 3Strategic CAD Labs, Intel Corporation, US Abstract High-speed I/O link is an important component in computer systems, and estimating its bit error rate (BER) is a critical task to guarantee its performance. In this paper, we propose an efficient method to estimate BER by Bayesian Model Fusion. Its key idea is to borrow conventional extrapolated BER value as prior knowledge, and combine it with additional measurement data to "calibrate" the BER value. This method can be viewed as an application of Bayesian Model Fusion (BMF) technique. We further propose some novel methodologies to make BMF applicable in the BER estimation case. In this way, we can sufficiently decrease the number of bits needed to estimate BER value. Several experiments demonstrate that our proposed method achieves up to 8x speed-up over direct estimation method. Download Paper (PDF; Only available from the DATE venue WiFi) |
| 17:30 | 8.6.2 | FAST DEPLOYMENT OF ALTERNATE ANALOG TEST USING BAYESIAN MODEL FUSION Speakers: John Liaperdos1, Haralampos Stratigopoulos2, Louay Abdallah2, Yiorgos Tsiatouhas3, Angela Arapoyanni4 and Xin Li5 1Technological Educational Institute of Peloponnese, GR; 2TIMA Laboratory, Université de Grenoble-Alpes/CNRS, FR; 3University of Ioannina, GR; 4National and Kapodistrian University of Athens, GR; 5Carnegie Mellon University, US Abstract In this paper, we address the problem of limited training sets for learning the regression functions in alternate analog test. Typically, a large volume of real data needs to be collected from different wafers and lots over a long period of time to be able to train the regression functions with accuracy across the whole design space and apply alternate test with high confidence. To avoid this delay and achieve a fast deployment of alternate test, we propose to use the Bayesian model fusion technique that leverages prior knowledge from simulation data and fuses this information with data from few real circuits to draw accurate regression functions across the whole design space. The technique is demonstrated for an alternate test designed for RF low noise amplifiers. Download Paper (PDF; Only available from the DATE venue WiFi) |
| 18:00 | 8.6.3 | BORDERSEARCH: AN ADAPTIVE IDENTIFICATION OF FAILURE REGIONS Speakers: Markus Dobler1, Manuel Harrant2, Monica Rafaila2, Georg Pelz2, Wolfgang Rosenstiel1 and Martin Bogdan3 1University of Tübingen, DE; 2Infineon Technologies, DE; 3University of Tübingen, Leipzig University, DE Abstract The reliability and safety of modern analog devices, e.g. in automotives, aircraft or consumer electronics, is influenced by many input parameters like supply voltage, ambient temperature or load resistances. In certain regions of this large parameter space, the device exhibits degraded performance or it fails completely. The validation of such a device has to find the regions of the input parameter space in which the device misbehaves. However, with several parameters, it is a complex task to determine these regions, especially if parameters interact. In this paper, we present the Bordersearch algorithm, which combines adaptive testing with a machine learning classifier to efficiently determine the border between passing and failing regions in the parameter space. Furthermore, this method enables sophisticated post-processing analysis, e.g. better visualizations and automatic ranking of the parameters according to their influence. This algorithm scales well to a high-dimensional parameter space and is robust against outliers and fuzzy borders. We show the effectiveness of this method on an automotive electro-mechanical system with eleven input parameters. Download Paper (PDF; Only available from the DATE venue WiFi) |
| 18:15 | 8.6.4 | A FAST SPATIAL VARIATION MODELING ALGORITHM FOR EFFICIENT TEST COST REDUCTION OF ANALOG/RF CIRCUITS Speakers: Hugo Gonçalves1, Xin Li2, Miguel Correia3, Vitor Tavares3, John Carulli4 and Kenneth Butler5 1CMU/FEUP, PT; 2Carnegie Mellon University, US; 3FEUP, PT; 4GLOBALFOUNDRIES, US; 5Texas Instruments, US Abstract In this paper, we adopt a novel numerical algorithm, referred to as dual augmented Lagrangian method (DALM), for efficient test cost reduction based on spatial variation modeling. The key idea of DALM is to derive the dual formulation of the L1-regularized least-squares problem posed by Virtual Probe (VP), which can be efficiently solved with substantially lower computational cost than its primal formulation. In addition, a number of unique properties associated with discrete cosine transform (DCT) are exploited to further reduce the computational cost of DALM. Our experimental results of an industrial RF transceiver demonstrate that the proposed DALM solver achieves up to 38x runtime speed-up over the conventional interior-point solver without sacrificing any performance on escape rate and yield loss for test applications. Download Paper (PDF; Only available from the DATE venue WiFi) |
| 18:30 | IP4-5, 656 | A HYBRID QUASI MONTE CARLO METHOD FOR YIELD AWARE ANALOG CIRCUIT SIZING TOOL Speakers: Engin Afacan, Günhan Dündar, Gonenc Berkol, Ali Emre Pusane and İsmail Faik Baskaya, Bogazici University, TR Abstract Efficient yield estimation methods are required by yield aware automatic sizing tools, where many iterative variability analyses are performed. Quasi Monte Carlo (QMC) is a popular approach, in which samples are generated more homogeneously, hence faster convergence is obtained compared to the conventional MC. However, since QMC is deterministic and has no natural variance, there is no convenient way to obtain estimation error bounds. To determine the confidence interval of the estimated yield, scrambled QMC, in which samples are randomly permuted, is run multiple times to obtain stochastic variance by sacrificing computational cost. To palliate this challenge, this paper proposes a hybrid method, where a single QMC is performed to determine infeasible solutions in terms of yield, which is followed by a few scrambled QMC analyses providing variance and confidence interval of the estimated yield. Yield optimization is performed considering the worst case of the current estimation, thus the optimizer guarantees that the solution will satisfy the confidence interval. Furthermore, a yield ranking mechanism is also developed to enforce the optimizer to search for more robust solutions. Download Paper (PDF; Only available from the DATE venue WiFi) |
| 18:31 | IP4-6, 179 | FEATURE SELECTION FOR ALTERNATE TEST USING WRAPPERS: APPLICATION TO AN RF LNA CASE STUDY Speakers: Manuel Barragan1 and Gildas Leger2 1TIMA Laboratory, FR; 2Instituto de Microelectronica de Sevilla, IMSE-CNM, (CSIC - Universidad de Sevilla), ES Abstract Testing analog, mixed-signal and RF circuits represents the main cost component for testing complex SoCs. A promising solution to alleviate this cost is the Alternate Test strategy. Alternate test is an indirect test approach that replaces costly specification measurements by simpler signatures. Machine learning techniques are then used to map signatures and performances. One key point that still remains as an open problem is the conception of adequate simple measurement candidates. This work presents efficient algorithms for selecting information rich signatures. Download Paper (PDF; Only available from the DATE venue WiFi) |
| 18:30 | End of session | |
| 19:30 | DATE Party in Museum of Grenoble (Musée de Grenoble, 5 Place de Lavalette, 38000 Grenoble, France)
As one of the main networking opportunities during the DATE week, the DATE Party states a perfect occasion to meet friends and colleagues in a relaxed atmosphere while enjoying local amenities. It will take place on March 11, 2015, from 19:30 to 23:00 in the renowned "Musée de Grenoble" (Grenoble Museum). This painting museum features a unique collection of ancient, modern and contemporary art including major masterpieces of classical Flemish, Dutch, Italian and Spanish painting and all the great pot-1945 contemporary art-trends, right up to the most recent artwork of the 2000s. During this evening, you can enjoy the famous French Cuisine and outstanding wines. Discover the region of the French Alps through ist cheese and wine specialties. The dinner will be accompanied by jazz songs and instrumental music from Anna Cruz and her vocal band. Another highlight will be the show waders "THE INSEPARABLES", sweet and ephemeral characters walking through the premises, releasing dreams and laughter. Furthermore, at the very beginning of the evening, from 20h00 to 21h30, you will have the opportunity to visit parts of the permanent collection of the museum (ninetieth and twenties century). Please kindly note that it is not a seated dinner. All delegates, exhibitors and their guests are invited to attend the party. Please be aware that entrance is only possible with a valid party ticket. Each full conference registration includes a ticket for the DATE Party (which needs to be booked during the online registration process though). Additional tickets can be purchased on-site at the registration desk (subject to availability of tickets). Price for extra ticket: 60 € per person. How to get there: The tram B has a stop called "Notre Dame Musee". That stop is next to the Museum. Attendees would take the tram A from Alpexpo and change for Tram B in one of the stations between "Gares" and "Maison du Tourisme" to get to the museum. The trip takes about 30 minutes. |
Visit us at DATE 2015
