Design and synthesis of efficient mac architectures for high speed decimal processor

Most of the commercial and financial data are stored in decimal fonn. Recently, support for decimal arithmetic has received increased attention due to the growing importance in financial analysis, banking, tax calculation, currency conversion, insurance, telephone billing and accounting. Performing decimal arithmetic with systems that do not support decimal computations may give a result with representation error, conversion error, and/or rounding error. In this world of precision, such errors are no more tolerable. The errors can be eliminated and better accuracy can be achieved if decimal computations are done using Decimal Floating Point (DFP) units. But the floating-point arithmetic units in today's general-purpose microprocessors are based on the binary number system, and the decimal computations are done using binary arithmetic. Only few common decimal numbers can be exactly represented in Binary Floating Point (BF P). ln many; cases, the law requires that results generated from financial calculations performed on a computer should exactly match with manual calculations. Currently many applications involving fractional decimal data perform decimal computations either in software or with a combination of software and hardware. The performance can be dramatically improved by complete hardware DFP units and this leads to the design of processors that include DF P hardware.VLSI implementations using same modular building blocks can decrease system design and manufacturing cost. A multiplexer realization is a natural choice from the viewpoint of cost and speed.This thesis focuses on the design and synthesis of efficient decimal MAC (Multiply ACeumulate) architecture for high speed decimal processors based on IEEE Standard for Floating-point Arithmetic (IEEE 754-2008). The research goal is to design and synthesize deeimal'MAC architectures to achieve higher performance.Efficient design methods and architectures are developed for a high performance DFP MAC unit as part of this research.

An efficient scheme for the evaluation of coding performance of high speed links incorporating joint error behaviour based on Monte Carlo Method

While channel coding is a standard method of improving a system’s energy efficiency in digital communications, its practice does not extend to high-speed links. Increasing demands in network speeds are placing a large burden on the energy efficiency of high-speed links and render the benefit of channel coding for these systems a timely subject. The low error rates of interest and the presence of residual intersymbol interference (ISI) caused by hardware constraints impede the analysis and simulation of coded high-speed links. Focusing on the residual ISI and combined noise as the dominant error mechanisms, this paper analyses error correlation through concepts of error region, channel signature, and correlation distance. This framework provides a deeper insight into joint error behaviours in high-speed links, extends the range of statistical simulation for coded high-speed links, and provides a case against the use of biased Monte Carlo methods in this setting

Simulation and Design of a high speed Solid State Switch for Excimer Laser

Excimerlaser sind gepulste Gaslaser, die Laseremission in Form von Linienstrahlung – abhängig von der Gasmischung – im UV erzeugen. Der erste entladungsgepumpte Excimerlaser wurde 1977 von Ischenko demonstriert. Alle kommerziell verfügbaren Excimerlaser sind entladungsgepumpte Systeme. Um eine Inversion der Besetzungsdichte zu erhalten, die notwendig ist, um den Laser zum Anschwingen zu bekommen, muss aufgrund der kurzen Wellenlänge sehr stark gepumpt werden. Diese Pumpleistung muss von einem Impulsleistungsmodul erzeugt werden. Als Schaltelement gebräuchlich sind Thyratrons, Niederdruckschaltröhren, deren Lebensdauer jedoch sehr limitiert ist. Deshalb haben sich seit Mitte der 1990iger Jahre Halbleiterschalter mit Pulskompressionsstufen auch in dieser Anwendung mehr und mehr durchgesetzt. In dieser Arbeit wird versucht, die Pulskompression durch einen direkt schaltenden Halbleiterstapel zu ersetzen und dadurch die Verluste zu reduzieren sowie den Aufwand für diese Pulskompression einzusparen. Zudem kann auch die maximal mögliche Repetitionsrate erhöht werden. Um die Belastung der Bauelemente zu berechnen, wurden für alle Komponenten möglichst einfache, aber leistungsfähige Modelle entwickelt. Da die normalerweise verfügbaren Daten der Bauelemente sich aber auf andere Applikationen beziehen, mussten für alle Bauteile grundlegende Messungen im Zeitbereich der späteren Applikation gemacht werden. Für die nichtlinearen Induktivitäten wurde ein einfaches Testverfahren entwickelt um die Verluste bei sehr hohen Magnetisierungsgeschwindigkeiten zu bestimmen. Diese Messungen sind die Grundlagen für das Modell, das im Wesentlichen eine stromabhängige Induktivität beschreibt. Dieses Modell wurde für den „magnetic assist“ benützt, der die Einschaltverluste in den Halbleitern reduziert. Die Impulskondensatoren wurden ebenfalls mit einem in der Arbeit entwickelten Verfahren nahe den späteren Einsatzparametern vermessen. Dabei zeigte sich, dass die sehr gebräuchlichen Class II Keramikkondensatoren für diese Anwendung nicht geeignet sind. In der Arbeit wurden deshalb Class I Hochspannungs- Vielschicht- Kondensatoren als Speicherbank verwendet, die ein deutlich besseres Verhalten zeigen. Die eingesetzten Halbleiterelemente wurden ebenfalls in einem Testverfahren nahe den späteren Einsatzparametern vermessen. Dabei zeigte sich, dass nur moderne Leistungs-MOSFET´s für diesen Einsatz geeignet sind. Bei den Dioden ergab sich, dass nur Siliziumkarbid (SiC) Schottky Dioden für die Applikation einsetzbar sind. Für die Anwendung sind prinzipiell verschiedene Topologien möglich. Bei näherer Betrachtung zeigt sich jedoch, dass nur die C-C Transfer Anordnung die gewünschten Ergebnisse liefern kann. Diese Topologie wurde realisiert. Sie besteht im Wesentlichen aus einer Speicherbank, die vom Netzteil aufgeladen wird. Aus dieser wird dann die Energie in den Laserkopf über den Schalter transferiert. Aufgrund der hohen Spannungen und Ströme müssen 24 Schaltelemente in Serie und je 4 parallel geschaltet werden. Die Ansteuerung der Schalter wird über hochisolierende „Gate“-Transformatoren erreicht. Es zeigte sich, dass eine sorgfältig ausgelegte dynamische und statische Spannungsteilung für einen sicheren Betrieb notwendig ist. In der Arbeit konnte ein Betrieb mit realer Laserkammer als Last bis 6 kHz realisiert werden, der nur durch die maximal mögliche Repetitionsrate der Laserkammer begrenzt war.

High-Speed Semiconductor Lasers based on Low-Dimensional Active Materials for Optical Telecommunication

The scope of this work is the fundamental growth, tailoring and characterization of self-organized indium arsenide quantum dots (QDs) and their exploitation as active region for diode lasers emitting in the 1.55 µm range. This wavelength regime is especially interesting for long-haul telecommunications as optical fibers made from silica glass have the lowest optical absorption. Molecular Beam Epitaxy is utilized as fabrication technique for the quantum dots and laser structures. The results presented in this thesis depict the first experimental work for which this reactor was used at the University of Kassel. Most research in the field of self-organized quantum dots has been conducted in the InAs/GaAs material system. It can be seen as the model system of self-organized quantum dots, but is not suitable for the targeted emission wavelength. Light emission from this system at 1.55 µm is hard to accomplish. To stay as close as possible to existing processing technology, the In(AlGa)As/InP (100) material system is deployed. Depending on the epitaxial growth technique and growth parameters this system has the drawback of producing a wide range of nano species besides quantum dots. Best known are the elongated quantum dashes (QDash). Such structures are preferentially formed, if InAs is deposited on InP. This is related to the low lattice-mismatch of 3.2 %, which is less than half of the value in the InAs/GaAs system. The task of creating round-shaped and uniform QDs is rendered more complex considering exchange effects of arsenic and phosphorus as well as anisotropic effects on the surface that do not need to be dealt with in the InAs/GaAs case. While QDash structures haven been studied fundamentally as well as in laser structures, they do not represent the theoretical ideal case of a zero-dimensional material. Creating round-shaped quantum dots on the InP(100) substrate remains a challenging task. Details of the self-organization process are still unknown and the formation of the QDs is not fully understood yet. In the course of the experimental work a novel growth concept was discovered and analyzed that eases the fabrication of QDs. It is based on different crystal growth and ad-atom diffusion processes under supply of different modifications of the arsenic atmosphere in the MBE reactor. The reactor is equipped with special valved cracking effusion cells for arsenic and phosphorus. It represents an all-solid source configuration that does not rely on toxic gas supply. The cracking effusion cell are able to create different species of arsenic and phosphorus. This constitutes the basis of the growth concept. With this method round-shaped QD ensembles with superior optical properties and record-low photoluminescence linewidth were achieved. By systematically varying the growth parameters and working out a detailed analysis of the experimental data a range of parameter values, for which the formation of QDs is favored, was found. A qualitative explanation of the formation characteristics based on the surface migration of In ad-atoms is developed. Such tailored QDs are finally implemented as active region in a self-designed diode laser structure. A basic characterization of the static and temperature-dependent properties was carried out. The QD lasers exceed a reference quantum well laser in terms of inversion conditions and temperature-dependent characteristics. Pulsed output powers of several hundred milli watt were measured at room temperature. In particular, the lasers feature a high modal gain that even allowed cw-emission at room temperature of a processed ridge wave guide device as short as 340 µm with output powers of 17 mW. Modulation experiments performed at the Israel Institute of Technology (Technion) showed a complex behavior of the QDs in the laser cavity. Despite the fact that the laser structure is not fully optimized for a high-speed device, data transmission capabilities of 15 Gb/s combined with low noise were achieved. To the best of the author`s knowledge, this renders the lasers the fastest QD devices operating at 1.55 µm. The thesis starts with an introductory chapter that pronounces the advantages of optical fiber communication in general. Chapter 2 will introduce the fundamental knowledge that is necessary to understand the importance of the active region`s dimensions for the performance of a diode laser. The novel growth concept and its experimental analysis are presented in chapter 3. Chapter 4 finally contains the work on diode lasers.

Development of directly modulated high speed telecommunication lasers based on surface defined feedback gratings

High-speed semiconductor lasers are an integral part in the implemen- tation of high-bit-rate optical communications systems. They are com- pact, rugged, reliable, long-lived, and relatively inexpensive sources of coherent light. Due to the very low attenuation window that exists in the silica based optical fiber at 1.55 μm and the zero dispersion point at 1.3 μm, they have become the mainstay of optical fiber com- munication systems. For the fabrication of lasers with gratings such as, distributed bragg reflector or distributed feedback lasers, etching is the most critical step. Etching defines the lateral dimmensions of the structure which determines the performance of optoelectronic devices. In this thesis studies and experiments were carried out about the exist- ing etching processes for InP and a novel dry etching process was de- veloped. The newly developed process was based on Cl2/CH4/H2/Ar chemistry and resulted in very smooth surfaces and vertical side walls. With this process the grating definition was significantly improved as compared to other technological developments in the respective field. A surface defined grating definition approach is used in this thesis work which does not require any re-growth steps and makes the whole fabrication process simpler and cost effective. Moreover, this grating fabrication process is fully compatible with nano-imprint lithography and can be used for high throughput low-cost manufacturing. With usual etching techniques reported before it is not possible to etch very deep because of aspect ratio dependent etching phenomenon where with increasing etch depth the etch rate slows down resulting in non-vertical side walls and footing effects. Although with our de- veloped process quite vertical side walls were achieved but footing was still a problem. To overcome the challenges related to grating defini- tion and deep etching, a completely new three step gas chopping dry etching process was developed. This was the very first time that a time multiplexed etching process for an InP based material system was demonstrated. The developed gas chopping process showed extra ordinary results including high mask selectivity of 15, moderate etch- ing rate, very vertical side walls and a record high aspect ratio of 41. Both the developed etching processes are completely compatible with nano imprint lithography and can be used for low-cost high-throughput fabrication. A large number of broad area laser, ridge waveguide laser, distributed feedback laser, distributed bragg reflector laser and coupled cavity in- jection grating lasers were fabricated using the developed one step etch- ing process. Very extensive characterization was done to optimize all the important design and fabrication parameters. The devices devel- oped have shown excellent performance with a very high side mode suppression ratio of more than 52 dB, an output power of 17 mW per facet, high efficiency of 0.15 W/A, stable operation over temperature and injected currents and a threshold current as low as 30 mA for almost 1 mm long device. A record high modulation bandwidth of 15 GHz with electron-photon resonance and open eye diagrams for 10 Gbps data transmission were also shown.

1.55 µm High-Speed QDs Lasers for Optical Telecommunication Applications

In this work investigation of the QDs formation and the fabrication of QD based semiconductor lasers for telecom applications are presented. InAs QDs grown on AlGaInAs lattice matched to InP substrates are used to fabricate lasers operating at 1.55 µm, which is the central wavelength for far distance data transmission. This wavelength is used due to its minimum attenuation in standard glass fibers. The incorporation of QDs in this material system is more complicated in comparison to InAs QDs in the GaAs system. Due to smaller lattice mismatch the formation of circular QDs, elongated QDs and quantum wires is possible. The influence of the different growth conditions, such as the growth temperature, beam equivalent pressure, amount of deposited material on the formation of the QDs is investigated. It was already demonstrated that the formation process of QDs can be changed by the arsenic species. The formation of more round shaped QDs was observed during the growth of QDs with As2, while for As4 dash-like QDs. In this work only As2 was used for the QD growth. Different growth parameters were investigated to optimize the optical properties, like photoluminescence linewidth, and to implement those QD ensembles into laser structures as active medium. By the implementation of those QDs into laser structures a full width at half maximum (FWHM) of 30 meV was achieved. Another part of the research includes the investigation of the influence of the layer design of lasers on its lasing properties. QD lasers were demonstrated with a modal gain of more than 10 cm-1 per QD layer. Another achievement is the large signal modulation with a maximum data rate of 15 Gbit/s. The implementation of optimized QDs in the laser structure allows to increase the modal gain up to 12 cm-1 per QD layer. A reduction of the waveguide layer thickness leads to a shorter transport time of the carriers into the active region and as a result a data rate up to 22 Gbit/s was achieved, which is so far the highest digital modulation rate obtained with any 1.55 µm QD laser. The implementation of etch stop layers into the laser structure provide the possibility to fabricate feedback gratings with well defined geometries for the realization of DFB lasers. These DFB lasers were fabricated by using a combination of dry and wet etching. Single mode operation at 1.55 µm with a high side mode suppression ratio of 50 dB was achieved.

Optimizing the Dynamic Distribution of Data-stream for High Speed Communications

The performances of high-speed network communications frequently rest with the distribution of data-stream. In this paper, a dynamic data-stream balancing architecture based on link information is introduced and discussed firstly. Then the algorithms for simultaneously acquiring the passing nodes and links of a path between any two source-destination nodes rapidly, as well as a dynamic data-stream distribution planning are proposed. Some related topics such as data fragment disposal, fair service, etc. are further studied and discussed. Besides, the performance and efficiency of proposed algorithms, especially for fair service and convergence, are evaluated through a demonstration with regard to the rate of bandwidth utilization. Hoping the discussion presented here can be helpful to application developers in selecting an effective strategy for planning the distribution of data-stream.

An event-based approach to validating solar wind speed predictions: high-speed enhancements in the Wang-Sheeley-Arge model

One of the primary goals of the Center for Integrated Space Weather Modeling (CISM) effort is to assess and improve prediction of the solar wind conditions in near‐Earth space, arising from both quasi‐steady and transient structures. We compare 8 years of L1 in situ observations to predictions of the solar wind speed made by the Wang‐Sheeley‐Arge (WSA) empirical model. The mean‐square error (MSE) between the observed and model predictions is used to reach a number of useful conclusions: there is no systematic lag in the WSA predictions, the MSE is found to be highest at solar minimum and lowest during the rise to solar maximum, and the optimal lead time for 1 AU solar wind speed predictions is found to be 3 days. However, MSE is shown to frequently be an inadequate “figure of merit” for assessing solar wind speed predictions. A complementary, event‐based analysis technique is developed in which high‐speed enhancements (HSEs) are systematically selected and associated from observed and model time series. WSA model is validated using comparisons of the number of hit, missed, and false HSEs, along with the timing and speed magnitude errors between the forecasted and observed events. Morphological differences between the different HSE populations are investigated to aid interpretation of the results and improvements to the model. Finally, by defining discrete events in the time series, model predictions from above and below the ecliptic plane can be used to estimate an uncertainty in the predicted HSE arrival times.

A practical method for testing high-speed networking hardware architectures

This paper deals with the key issues encountered in testing during the development of high-speed networking hardware systems by documenting a practical method for "real-life like" testing. The proposed method is empowered by modern and commonly available Field Programmable Gate Array (FPGA) technology. Innovative application of standard FPGA blocks in combination with reconfigurability are used as a back-bone of the method. A detailed elaboration of the method is given so as to serve as a general reference. The method is fully characterised and compared to alternatives through a case study proving it to be the most efficient and effective one at a reasonable cost.

Multivariable cluster analysis for high-speed industrial machinery

The overall operation and internal complexity of a particular production machinery can be depicted in terms of clusters of multidimensional points which describe the process states, the value in each point dimension representing a measured variable from the machinery. The paper describes a new cluster analysis technique for use with manufacturing processes, to illustrate how machine behaviour can be categorised and how regions of good and poor machine behaviour can be identified. The cluster algorithm presented is the novel mean-tracking algorithm, capable of locating N-dimensional clusters in a large data space in which a considerable amount of noise is present. Implementation of the algorithm on a real-world high-speed machinery application is described, with clusters being formed from machinery data to indicate machinery error regions and error-free regions. This analysis is seen to provide a promising step ahead in the field of multivariable control of manufacturing systems.

Detection of faults in a high speed packaging machine using a multilayer perceptron (MLP)

In high speed manufacturing systems, continuous operation is desirable, with minimal disruption for repairs and service. An intelligent diagnostic monitoring system, designed to detect developing faults before catastrophic failure, or prior to undesirable reduction in output quality, is a good means of achieving this. Artificial neural networks have already been found to be of value in fault diagnosis of machinery. The aim here is to provide a system capable of detecting a number of faults, in order that maintenance can be scheduled in advance of sudden failure, and to reduce the necessity to replace parts at intervals based on mean time between failures. Instead, parts will need to be replaced only when necessary. Analysis of control information in the form of position error data from two servomotors is described.