Cette thèse concerne l'analyse de défaillances de circuits VLSI et plus particulièrement la détection de défauts sur des circuits (microprocesseurs) à structure non connue. Une méthodologie basée sur balayage fonctionnant en contraste de potentiel est proposée. Les différents outils nécessaires à la mise en œuvre de cette méthodologie sont ensuite développés. les principaux résultats obtenus sont exposés, résultats permettant de démontrer la faisabilité de cette méthodologie. Une deuxième partie décrit un dispositif original de formation et de mémorisation d'images à semi-conducteur réalisable en technologie MOS. Les principales caractéristiques de ce capteur sont présentées ainsi que les résultats de mesures effectuées sur un circuit prototype. Enfin des améliorations de ce dispositif sont proposés

La première étape de l'étude se traduit par la proposition d'utilisation de l'ABC 90 comme organe d'allocation de bus dans différentes configurations d'architectures et ce par adjonction d'éléments discrets. La seconde étape consiste à proposer un circuit intégré d'arbitre de bus multiprotocole en partant des spécifications de l'ABC 90 et en y intégrant les résultats obtenus dans la proposition précédente. La validation de ces deux propositions a été obtenue par simulation.

L'accroissement de la densité d'intégration des circuits intégrés exige des moyens de contrôle d'une extrème précision. La microscopie électronique à balayage en contraste de potentiel convient particulièrement. La thèse présente tous les aspects: organisation, observabilité, méthodes d'observation, modes de traitement et les applications à la mise au point de circuits prototypes, l'analyse des défaillances, le contrôle de qualité, la recherche des limites de fonctionnement, la restructuration. Discussion de la nécessité du développement de méthodologies d'utilisation

This work presents a new energy saving technique for modern digital designs. We propose Time Interleaved Multi-Rail (TIMR) - a method for providing two dynamic supply rails to a circuit. This technique uses the first supply rail to mask the transition delay while changing the voltage of the second rail. We examine the design of TIMR as well as the implementation and considerations. We propose a number of control schemes that range from traditional DVFS to "race to sleep". This thesis also shows simulations of the technique using a existing voltage regulator in order to find the time and energy overhead of implementing the design. We find a 100μs switching time delay and 118μJ energy overhead associated with changing the voltage rail. This work concludes with comparisons to current energy saving techniques. / Graduation date: 2012

The last few decades have been a very exciting period in thedevelopment of micro-electronics and brought us to the brink ofimplementing entire systems on a single chip, on a hithertounimagined scale. However an unforeseen challenge has croppedup in the form of managing wires, which have become the mainbottleneck in performance, masking the blinding speed of activedevices. A major problem is that increasingly complicatedeffects need to be modelled, but the computational complexityof any proposed model needs to be low enough to allow manyiterations in a design cycle. This thesis addresses the issue of closed form modelling ofthe response of coupled interconnect systems. Following astrict mathematical approach, second order models for thetransfer functions of coupled RC trees based on the first andsecond moments of the impulse response are developed. The2-pole-1-zero transfer function that is the best possible fromthe available information is obtained for the signal path fromeach driver to the output in multiple aggressor systems. Thisallows the complete response to be estimated accurately bysumming up the individual waveforms. The model represents theminimum complexity for a 2-pole-1-zero estimate, for this classof circuits. Also proposed are new techniques for the optimisation ofwires in on-chip buses. Rather than minimising the delay overeach individual wire, the configuration that maximises thetotal bandwidth over a number of parallel wires isinvestigated. It is shown from simulations that there is aunique optimal solution which does not necessarily translate tothe maximum possible number of wires, and in fact deviatesconsiderably from it when the resources available for repeatersare limited. Analytic guidelines dependent only on processparameters are derived for optimal sizing of wires andrepeaters. Finally regular tiled architectures with a commoncommunication backplane are being proposed as being the mostefficient way to implement systems-on-chip in the deepsubmicron regime. This thesis also considers the feasibility ofimplementing a regular packet-switched network-on-chip in atypical future deep submicron technology. All major physicalissues and challenges are discussed for two differentarchitectures and important limitations are identified.

Process variability is the pivotal factor impacting the design of high yield integrated circuits and systems in deep sub-micron CMOS technologies. The electrical and physical properties of transistors and interconnects, the building blocks of integrated circuits, are prone to significant variations that directly impact the performance and power consumption of the fabricated devices, severely impacting the manufacturing yield. However, the large number of the transistors on a single chip adds even more challenges for the analysis of the variation effects, a critical task in diagnosing the cause of failure and designing for yield. Reliable and efficient statistical analysis methodologies in various design phases are key to predict the yield before entering such an expensive fabrication process. In this thesis, the impacts of process variations are examined at three different levels: device, circuit, and micro-architecture. The variation models are provided for each level of abstraction, and new methodologies are proposed for efficient statistical analysis and design under variation. At the circuit level, the variability analysis of three crucial sub-blocks of today's system-on-chips, namely, digital circuits, memory cells, and analog blocks, are targeted. The accurate and efficient yield analysis of circuits is recognized as an extremely challenging task within the electronic design automation community. The large scale of the digital circuits, the extremely high yield requirement for memory cells, and the time-consuming analog circuit simulation are major concerns in the development of any statistical analysis technique. In this thesis, several sampling-based methods have been proposed for these three types of circuits to significantly improve the run-time of the traditional Monte Carlo method, without compromising accuracy. The proposed sampling-based yield analysis methods benefit from the very appealing feature of the MC method, that is, the capability to consider any complex circuit model. However, through the use and engineering of advanced variance reduction and sampling methods, ultra-fast yield estimation solutions are provided for different types of VLSI circuits. Such methods include control variate, importance sampling, correlation-controlled Latin Hypercube Sampling, and Quasi Monte Carlo. At the device level, a methodology is proposed which introduces a variation-aware design perspective for designing MOS devices in aggressively scaled geometries. The method introduces a yield measure at the device level which targets the saturation and leakage currents of an MOS transistor. A statistical method is developed to optimize the advanced doping profiles and geometry features of a device for achieving a maximum device-level yield. Finally, a statistical thermal analysis framework is proposed. It accounts for the process and thermal variations simultaneously, at the micro-architectural level. The analyzer is developed, based on the fact that the process variations lead to uncertain leakage power sources, so that the thermal profile, itself, would have a probabilistic nature. Therefore, by a co-process-thermal-leakage analysis, a more reliable full-chip statistical leakage power yield is calculated.

Variability has become one of the vital challenges that the designers of integrated circuits encounter. variability becomes increasingly important. Imperfect manufacturing process manifest itself as variations in the design parameters. These variations and those in the operating environment of VLSI circuits result in unexpected changes in the timing, power, and reliability of the circuits. With scaling transistor dimensions, process and environmental variations become significantly important in the modern VLSI design. A smaller feature size means that the physical characteristics of a device are more prone to these unaccounted-for changes. To achieve a robust design, the random and systematic fluctuations in the manufacturing process and the variations in the environmental parameters should be analyzed and the impact on the parametric yield should be addressed. This thesis studies the challenges and comprises solutions for designing robust VLSI systems in the presence of variations. Initially, to get some insight into the system design under variability, the parametric yield is examined for a small circuit. Understanding the impact of variations on the yield at the circuit level is vital to accurately estimate and optimize the yield at the system granularity. Motivated by the observations and results, found at the circuit level, statistical analyses are performed, and solutions are proposed, at the system level of abstraction, to reduce the impact of the variations and increase the parametric yield. At the circuit level, the impact of the supply and threshold voltage variations on the parametric yield is discussed. Here, a design centering methodology is proposed to maximize the parametric yield and optimize the power-performance trade-off under variations. In addition, the scaling trend in the yield loss is studied. Also, some considerations for design centering in the current and future CMOS technologies are explored. The investigation, at the circuit level, suggests that the operating temperature significantly affects the parametric yield. In addition, the yield is very sensitive to the magnitude of the variations in supply and threshold voltage. Therefore, the spatial variations in process and environmental variations make it necessary to analyze the yield at a higher granularity. Here, temperature and voltage variations are mapped across the chip to accurately estimate the yield loss at the system level. At the system level, initially the impact of process-induced temperature variations on the power grid design is analyzed. Also, an efficient verification method is provided that ensures the robustness of the power grid in the presence of variations. Then, a statistical analysis of the timing yield is conducted, by taking into account both the process and environmental variations. By considering the statistical profile of the temperature and supply voltage, the process variations are mapped to the delay variations across a die. This ensures an accurate estimation of the timing yield. In addition, a method is proposed to accurately estimate the power yield considering process-induced temperature and supply voltage variations. This helps check the robustness of the circuits early in the design process. Lastly, design solutions are presented to reduce the power consumption and increase the timing yield under the variations. In the first solution, a guideline for floorplaning optimization in the presence of temperature variations is offered. Non-uniformity in the thermal profiles of integrated circuits is an issue that impacts the parametric yield and threatens chip reliability. Therefore, the correlation between the total power consumption and the temperature variations across a chip is examined. As a result, floorplanning guidelines are proposed that uses the correlation to efficiently optimize the chip's total power and takes into account the thermal uniformity. The second design solution provides an optimization methodology for assigning the power supply pads across the chip for maximizing the timing yield. A mixed-integer nonlinear programming (MINLP) optimization problem, subject to voltage drop and current constraint, is efficiently solved to find the optimum number and location of the pads.

Finite impulse response (FIR) digital filters are frequently used in many digital signal processing and communication applications, such as IS-95 CDMA, Digital Mobile Phone Systems (D-AMPS), etc. FIR filter achieves the frequency response of system requirement using a series of multiplications and additions. Previous papers on FIR hardware implementations usually focus on reducing area and delay of the multiple constant multiplications (MCM) through common sub-expression elimination (CSE) in the transpose FIR filter structure. In this thesis, we first perform optimization for the quantization of FIR filter coefficients that satisfy the target frequency response. Then suitable encoding methods are adopted to reduce the height of the partial products of the MCM in the direct FIR filter structure. Finally, by jointly considering the errors in the truncated multiplications and additions, we can design the hardware-efficient FIR filter that meets the bit accuracy requirement. Experimental results show that although CSE in the transpose FIR structure can reduce more area in MCM, the direct form takes smaller area in registers. Compared with previous approaches, the proposed FIR implementations with direct form has the minimum area cost.

This work is a first step towards a long-term goal of understanding computations occurring in the brain and using those principles to make more efficient machines. The traditional computing paradigm calls for using digital supercomputers to simulate large scale brain-like neural networks resulting in large power consumption which limits scalability or model detail. For example, IBM's digital simulation of a cat brain with simplistic neurons and synapses consumes power equivalent to that of a thousand houses! Instead of digital methods, this work uses analog processing concepts to develop scalable, low-power silicon models of neurons which have been shown to be around ten thousand times more power efficient. This has been achieved by modeling the dynamical behavior of Hodgkin-Huxley (H-H) or Morris-Lecar type equations instead of modeling the exact equations themselves. In particular, the two silicon neuron designs described exhibit a Hopf and a saddle-node bifurcation. Conditions for the bifurcations allow the identification of correct biasing regimes for the neurons. Also, since the hardware neurons compute in real time, they can be used for dynamic clamp protocols in addition to computational experiments. To empower this analog implementation with the flexibility of a digital simulation, a family of field programmable analog array (FPAA) architectures have been developed in 0.35 um CMOS that provide reconfigurability in the network of neurons as well as tunability of individual neuron parameters. This programmability is obtained using floating-gate (FG) transistors. The neurons are organized in blocks called computational analog blocks (CAB) which are embedded in a programmable switch matrix. An unique feature of the architecture is that the switches, being FG elements, can be used also for computation leading to more than 50,000 analog parameters in 9 sq. mm. Several neural systems including central pattern generators and coincidence detectors are demonstrated. Also, a separate chip that is capable of implementing signal processing algorithms has been designed by modifying the CAB elements to include transconductors, multipliers etc. Several systems including an AM demodulator and a speech processor are presented. An important contribution of this work is developing an architecture for programming the FG elements over a wide dynamic range of currents. An adaptive logarithmic transimpedance amplifier is used for this purpose. This design provides a general solution for wide dynamic range current measurement with a low power dissipation and has been used in imaging chips too. A new generation of integrated circuits have also been designed that are 25 sq. mm in area and contain several new features including adaptive synapses and support for smart sensors. These designs and the previous ones should allow prototyping and rapid development of several neurally inspired systems and pave the path for the design of larger and more complex brain like adaptive neural networks.

Power has become an important design closure parameter in today’s ultra low submicron digital designs. The impact of the increase in power is multi-discipline to researchers ranging from power supply design, power converters or voltage regulators design, system, board and package thermal analysis, power grid design and signal integrity analysis to minimizing power itself. This work focuses on challenges arising due to increase in power to power grid design and analysis. Challenges arising due to lower geometries and higher power are very well researched topics and there is still lot of scope to continue work. Traditionally, designs go through average IR drop analysis. Average IR drop analysis is highly dependent on current dissipation estimation. This work proposes a vector less probabilistic toggle estimation which is extension of one of the approaches proposed in literature. We have further used toggles computed using this approach to estimate power of ISCAS89 benchmark circuits. This provides insight into quality of toggles being generated. Power Estimation work is further extended to comprehend with various state of the art methodologies available i.e. spice based power estimation, logic simulation based power estimation, commercially available tool comparisons etc. We finally arrived at optimum flow recommendation which can be used as per design need and schedule. Today’s design complexity – high frequencies, high logic densities and multiple level clock and power gating - has forced design community to look beyond average IR drop. High rate of switching activities induce power supply fluctuations to cells in design which is known as instantaneous IR drop. However, there is no good analysis methodology in place to analyze this phenomenon. Ad hoc decoupling planning and on chip intrinsic decoupling capacitance helps to contain this noise but there is no guarantee. This work also applies average toggle computation approach to compute instantaneous IR drop analysis for designs. Instantaneous IR drop is also known as dynamic IR drop or power supply noise. We are proposing cell characterization methodology for standard cells. This data is used to build power grid model of the design. Finally, the power network is solved to compute instantaneous IR drop. Leakage Power Minimization has forced design teams to do complex power gating – multilevel MTCMOS usage in Power Grid. This puts additonal analysis challenge for Power Grid in terms of ON/OFF sequencing and noise injection due to it. This work explains the state of art here and highlights some of the issues and trade offs using MTCMOS logic. It further suggests a simple approach to quickly access the impact of MTCMOS gates in Power Grid in terms of peak currents and IR drop. Alternatively, the approach suggested also helps in MTCMOS gate optimization. Early leakage optimization overhead can be computed using this approach.