L'étude des principaux domaines d'application et des techniques essentielles du traitement d'images permet de dégager des interrogations directes pour le concepteur d'architectures d'ordinateurs. Quelques unes des voies possibles pour leurs solutions matérielles sont ensuite présentées dans leurs principes et a l'aide d'exemples. Présentation de deux études : la machine ROMULARD (multi-microprocesseur) pour la saisie et le traitement d'images et le systeme KIDS, architecture plus ambitieuse alliant les aspects logiciels et matériels.

Des études récentes montrent que le modèle de collage logique ne convient pas pour représenter les défauts réels qui peuvent survenir dans les circuits intégrés. C'est pourquoi on recherche des méthodes de test basées sur des hypothèses de pannes analytiques. Problème de la conception des circuits autotestables vis-a-vis d'hypothèses de pannes analytiques: méthodes et règles générales pour les circuits fonctionnels N-MOS "fortement garantis contre les fautes" ("strongly fault secure": sfs). Nouveaux codes. Classe des contrôleurs " à codes fortement disjoints" (" strongly code disjoint": scd). Application des méthodes à l'étude d'un microprocesseur mc68000 autotestable.

La probabilité des fautes transitoires augmente avec l'évolution des technologies. Plusieurs approches ont été proposées pour analyser très tôt l'impact de ces fautes sur un circuit numérique. Il est notamment possible d'utiliser une approche fondée sur l'injection de fautes dans une description VHDL au niveau RTL. Dans cette thèse, nous apportons plusieurs contributions à ce type d'analyse. Un premier aspect considéré est la prise en compte de l'environnement du circuit numérique lors des campagnes d'injection. Ainsi, une approche basée sur une analyse de sûreté de fonctionnement multi-niveaux a été développée et appliquée sur un exemple. Les injections sont réalisées dans le circuit numérique décrit au niveau RTL alors que le reste du système est décrit à un niveau d'abstraction plus élevé. L'analyse des résultats montre que certaines défaillances apparaissant au niveau du circuit n'ont en fait aucun impact sur le système. Nous présentons ensuite les avantages de la combinaison de deux types d'analyses : la classification des fautes en fonction de leurs effets, et l'analyse plus détaillée des configurations d'erreurs activées dans le circuit. Une campagne d'injection de fautes de type SEU a été réalisée sur un microcontrôleur 8051 décrit au niveau RTL. Les résultats montrent que la combinaison des analyses permet au concepteur de localiser les points critiques, facilitant l'étape de durcissement. Ils montrent également que, dans le cas d'un processeur à usage général, les configurations d'erreurs peuvent être dépendantes du programme exécuté. Cette étude a également permis de montrer que l'injection d'un très faible pourcentage des fautes possibles permet déjà d'obtenir des informations utiles pour le concepteur. La même méthodologie a été utilisée pour valider la robustesse obtenue avec un durcissement au niveau logiciel. Les résultats montrent que certaines fautes ne sont pas détectées par les mécanismes implémentés bien que ceux-ci aient été préalablement validés par des injections de fautes basées sur un simulateur de jeu d'instructions. Le dernier aspect de cette thèse concerne l'injection de fautes dans des blocs analogiques. En fait très peu de travaux traitent du sujet. Nous proposons donc un flot global d'analyse pour circuits numériques, analogiques ou mixtes, décrits au niveau comportemental. La possibilité d'injecter des fautes dans des blocs analogiques est discutée. Les résultats obtenus sur une PLL, choisie comme cas d'étude, sont analysés et montrent la faisabilité de l'injection de fautes dans des blocs analogiques. Pour valider le flot, des injections de fautes sont également réalisées au niveau transistor et comparées à celles réalisées à haut niveau. Il apparaît une bonne corrélation entre les résultats obtenus aux deux niveaux.

<p>Multiplikatorer används ofta som ett byggblock vid konstruktion av kretsar som digitala filter, FFT-processorer och aritmetiska enheter. Olika trädstrukturer används i"höghastighet"applikationer för multiplikatorer. En typ av träd,"overturned-stairs"(OS) som är ett adderarträd av första ordningen har uppvisat lika optimal prestanda med avseende på hastighet som Wallace-träd, vid 18 eller färre ingångar. I moderna integrerade kretsar, ger ledningar och kopplingar upphov till fördröjningar och parasitiska laster. I en jämförelse mellan Wallace-träd, och OS1-träd har det sistnämda kortare och mindre komplicerad ledningsdragningar och är därför mer ändamålsenlig för VLSI implementationer.</p>

<p>A fast and energy-efficient multiplier is always needed in electronics industry especially DSP, image processing and arithmetic units in microprocessors. Multiplier is such an important element which contributes substantially to the total power consumption of the system. On VLSI level, the area also becomes quite important as more area means more system cost. Speed is another key parameter while designing a multiplier for a specific application.</p><p>These three parameters i.e. power, area and speed are always traded off. Speaking of DSP processors, area and speed of MAC unit are the most important factors. But sometimes, increasing speed also increases the power consumption, so there is an upper bound of speed for a given power criteria. Considering the battery operated portable multimedia devices, low power and fast designs of multipliers are more important than area.</p><p>The design of a low power, high speed and area efficient multiplier is thus the goal of my thesis work. The projected plan is to instantiate a good design and modify it for low power and speed and prepare its layout using 90nm technology in Cadence®. For that purpose study has been performed on a number of research papers presented in section 7 and selected one of the architecture presented by Jung-Yup Kang and Jean-Luc Gaudiot. They presented a unique technique for power reduction in Wallace tree multipliers. They have proposed a method to calculate 2’s complement of multiplicand for final Partial Product Row (PPR) if using MBE technique. This method has been used in the design for speed enhancement and power reduction.</p><p>The ultimate purpose is to come up with such an architecture which is energy and area efficient than a conventional multiplier at the same performance level. This report describes the design and evaluation of new energy-efficient 32-bit multiplier architecture by comparing its power, performance and chip area to those of a conventional 32-bit multiplier. The report throws light on the basic principles and methods of binary multiplication process and also the power consumption issues related to multipliers. The new algorithm, which reduces the last negative signal in the partial product row is discussed to develop the new architecture. A power performance comparison is shown. The simulation results show that the new architecture is 46 % energy-efficient than a conventional multiplier at the same performance level. The number of transistors used is 34% less and also it consumes 25% less chip area in 90nm CMOS technology.</p>

In wireless communications, the use of multiple antennas at both the transmitter and the receiver is a key technology to enable high data rate transmission without additional bandwidth or transmit power. Multiple-input multiple-output (MIMO) schemes are widely used in many wireless standards, allowing higher throughput using spatial multiplexing techniques. MIMO soft detection poses significant challenges to the MIMO receiver design as the detection complexity increases exponentially with the number of antennas. As the next generation wireless system is pushing for multi-Gbps data rate, there is a great need for high-throughput low-complexity soft-output MIMO detector. The brute-force implementation of the optimal MIMO detection algorithm would consume enormous power and is not feasible for the current technology. We propose a reduced-complexity soft-output MIMO detector architecture based on a trellis-search method. We convert the MIMO detection problem into a shortest path problem. We introduce a path reduction and a path extension algorithm to reduce the search complexity while still maintaining sufficient soft information values for the detection. We avoid the missing counter-hypothesis problem by keeping multiple paths during the trellis search process. The proposed trellis-search algorithm is a data-parallel algorithm and is very suitable for high speed VLSI implementation. Compared with the conventional tree-search based detectors, the proposed trellis-based detector has a significant improvement in terms of detection throughput and area efficiency. The proposed MIMO detector has great potential to be applied for the next generation Gbps wireless systems by achieving very high throughput and good error performance. The soft information generated by the MIMO detector will be processed by a channel decoder, e.g. a low-density parity-check (LDPC) decoder or a Turbo decoder, to recover the original information bits. Channel decoder is another very computational-intensive block in a MIMO receiver SoC (system-on-chip). We will present high-performance LDPC decoder architectures and Turbo decoder architectures to achieve 1+ Gbps data rate. Further, a configurable decoder architecture that can be dynamically reconfigured to support both LDPC codes and Turbo codes is developed to support multiple 3G/4G wireless standards. We will present ASIC and FPGA implementation results of various MIMO detectors, LDPC decoders, and Turbo decoders. We will discuss in details the computational complexity and the throughput performance of these detectors and decoders.

The power consumption in commercial processors and application specific integrated circuits increases with decreasing technology nodes. Power saving techniques have become a first class design point for current and future VLSI systems. These systems employ large on-chip SRAM memories. Reducing memory leakage power while maintaining data integrity is a key criterion for modern day systems. Unfortunately, state of the art techniques like power-gating can only be applied to logic as these would destroy the contents of the memory if applied to a SRAM system. Fortunately, previous works have noted large temporal and spatial locality for data patterns in commerical processors as well as application specific ICs that work on images, audio and video data. This thesis presents a novel column based Energy Compression technique that saves SRAM power by selectively turning off cells based on a data pattern. This technique is applied to study the power savings in application specific inegrated circuit SRAM memories and can also be applied for commercial processors. The thesis also evaluates the effects of processing images before storage and data cluster patterns for optimizing power savings.

No description available.

Multiplikatorer används ofta som ett byggblock vid konstruktion av kretsar som digitala filter, FFT-processorer och aritmetiska enheter. Olika trädstrukturer används i"höghastighet"applikationer för multiplikatorer. En typ av träd,"overturned-stairs"(OS) som är ett adderarträd av första ordningen har uppvisat lika optimal prestanda med avseende på hastighet som Wallace-träd, vid 18 eller färre ingångar. I moderna integrerade kretsar, ger ledningar och kopplingar upphov till fördröjningar och parasitiska laster. I en jämförelse mellan Wallace-träd, och OS1-träd har det sistnämda kortare och mindre komplicerad ledningsdragningar och är därför mer ändamålsenlig för VLSI implementationer.

A fast and energy-efficient multiplier is always needed in electronics industry especially DSP, image processing and arithmetic units in microprocessors. Multiplier is such an important element which contributes substantially to the total power consumption of the system. On VLSI level, the area also becomes quite important as more area means more system cost. Speed is another key parameter while designing a multiplier for a specific application. These three parameters i.e. power, area and speed are always traded off. Speaking of DSP processors, area and speed of MAC unit are the most important factors. But sometimes, increasing speed also increases the power consumption, so there is an upper bound of speed for a given power criteria. Considering the battery operated portable multimedia devices, low power and fast designs of multipliers are more important than area. The design of a low power, high speed and area efficient multiplier is thus the goal of my thesis work. The projected plan is to instantiate a good design and modify it for low power and speed and prepare its layout using 90nm technology in Cadence®. For that purpose study has been performed on a number of research papers presented in section 7 and selected one of the architecture presented by Jung-Yup Kang and Jean-Luc Gaudiot. They presented a unique technique for power reduction in Wallace tree multipliers. They have proposed a method to calculate 2’s complement of multiplicand for final Partial Product Row (PPR) if using MBE technique. This method has been used in the design for speed enhancement and power reduction. The ultimate purpose is to come up with such an architecture which is energy and area efficient than a conventional multiplier at the same performance level. This report describes the design and evaluation of new energy-efficient 32-bit multiplier architecture by comparing its power, performance and chip area to those of a conventional 32-bit multiplier. The report throws light on the basic principles and methods of binary multiplication process and also the power consumption issues related to multipliers. The new algorithm, which reduces the last negative signal in the partial product row is discussed to develop the new architecture. A power performance comparison is shown. The simulation results show that the new architecture is 46 % energy-efficient than a conventional multiplier at the same performance level. The number of transistors used is 34% less and also it consumes 25% less chip area in 90nm CMOS technology.