Georges Gielen

This book describes new tools for front end analog designers, starting with global variation-aware sizing, and extending to novel variation-aware topology design. The tools aid design through automation, but more importantly, they also aid designer insight through automation. We now describe four design tasks, each more general than the previous, and how this book contributes design aids and insight aids to each. The ?rst designer task targeted is global robust sizing. This task is supported by a design tool that does automated, globally reliable, variation-aware s- ing (SANGRIA),and an insight-aiding tool that extracts designer-interpretable whitebox models that relate sizings to circuit performance (CAFFEINE). SANGRIA searches on several levels of problem dif?culty simultaneously, from lower cheap-to-evaluate “exploration” layers to higher full-evaluation “exploitation” layers (structural homotopy). SANGRIAmakes maximal use of circuit simulations by performing scalable data mining on simulation results to choose new candidate designs. CAFFEINE accomplishes its task by tre- ing function induction as a tree-search problem. It constrains its tree search space via a canonical-functional-form grammar, and searches the space with grammatically constrained genetic programming. The second designer task is topology selection/topology design. Topology selection tools must consider a broad variety of topologies such that an app- priate topology is selected, must easily adapt to new semiconductor process nodes, and readily incorporate new topologies. Topology design tools must allow designers to creatively explore new topology ideas as rapidly as possible.

nalog circuits are fascinating artifacts. They manipulate signals whose informa- Ationcontentisrichcomparedtodigitalsignalsthatcarryminimalamountofinf- mation;theyaredelicateinthatanyperturbationduetoparasiticelements,todelays,to interactionswithotherelementsandwiththeenvironmentmaycauseasigni?cantloss ofinformation. Thedif?cultyindealingwiththeseartifactsistoprotectthemfromall possibleattacks, evenminorones, fromthephysicalworld. Theironyisthattheyare oftenusedtofunnelinformationfromandtothephysicalworldtoandfromtheabstr- tionofthedigitalworldandforthisfunction, theyareirreplaceable. Nowonderthen that analog designers form a club of extraordinary gentlemen where art (or magic?) ratherthanscienceisthesharedtrade. Theyaredif?culttotrainsinceexperienceand intuitionarethetraitsthat characterize them. Andthey have dif?cultiesinexplaining what is the process they use to reach satisfactory results. Tools used for design (s- ulation) are mainly replacing the test benches of an experimental lab. However, the growing complexity of the integrated systems being designed today together with the increasing fragility of analog components brought about by shrinking geometries and reducedpowerconsumptionisposingseverechallengestotraditionalanalogdesigners to produce satisfactory results in a short time. At the same time, the need for expe- enced analog designers has increased constantly since almost all designs, because of integration,docontainanalogcomponents. Thissituationhascreatedastronginterest in developing design methodologies and supporting tools that are based on rigorous, mathematically literate, approaches. Doing so will make it possible to leverage the expertiseofseasonedanalogdesignersandtotrainnewgenerationsfasterandbetter. Inthepast, severalattemptshavebeenmadeinacademia andindustrytocreatethese methodologies and to extend the set of tools available. They have had questionable acceptance in the analog design community. However, recently, a ?urry of start-ups andincreasedinvestmentbyEDAcompaniesinnoveltoolssignalasigni?cantchange inmarketattentiontotheanalogdomain. Ipersonallybelievethattosubstantially- prove quality and design time, tools are simply insuf?cient. A design methodology based on a hierarchy of abstraction layers, successive re?nement between two ad- cent layers, and extensive veri?cation at every layer is necessary. To do so, we need to build theories and models that have strong mathematical foundations. The analog design technology community is as strong as it has ever been.

As the miniaturization of semiconductor technology continues, electronic s- tems on chips o?er a more extensive and more complex functionality with better performance, higher frequencies and less power consumption. Whereas digital designers can take full advantage of the availability of design auto- tiontoolstobuildhugesystems,thelackofsupportbycomputerprogramsfor di?erent abstraction levels makes analog design a time-consuming handcraft which limits the possibilities to implement large systems. Various approaches for?ndingoptimalvaluesfortheparametersofanalogcells,likeopamps,have been investigated since the mid-1980s, and they have made their entrance in commercial applications. However, a larger impact on the performance is - pected if tools are developed which operate on a higher abstraction level and consider multiple architectural choices to realize a particular functionality. In this book, the opportunities, conditions, problems, solutions and systematic methodologies for this new generation of analog CAD tools are examined. Theoutlineofthisbookisasfollows. Inthe?rstpart,thecharacteristicsof the analog design process are systematically analyzed and several approaches for automated analog synthesis are summarized. Comparison of their prop- ties with the requirements for high-level synthesis of analog and mixed-signal systems results in a new design paradigm: the high-level design ?ow based on generic behavior. This design approach involves a modeling strategy using generic behavioral models and a synthesis strategy leading to the exploration of a heterogeneous design space containing di?erent architectures. The modeling strategy is further elaborated in Part II.

This book describes new tools for front end analog designers, starting with global variation-aware sizing, and extending to novel variation-aware topology design. The tools aid design through automation, but more importantly, they also aid designer insight through automation. We now describe four design tasks, each more general than the previous, and how this book contributes design aids and insight aids to each. The ?rst designer task targeted is global robust sizing. This task is supported by a design tool that does automated, globally reliable, variation-aware s- ing (SANGRIA),and an insight-aiding tool that extracts designer-interpretable whitebox models that relate sizings to circuit performance (CAFFEINE). SANGRIA searches on several levels of problem dif?culty simultaneously, from lower cheap-to-evaluate “exploration” layers to higher full-evaluation “exploitation” layers (structural homotopy). SANGRIAmakes maximal use of circuit simulations by performing scalable data mining on simulation results to choose new candidate designs. CAFFEINE accomplishes its task by tre- ing function induction as a tree-search problem. It constrains its tree search space via a canonical-functional-form grammar, and searches the space with grammatically constrained genetic programming. The second designer task is topology selection/topology design. Topology selection tools must consider a broad variety of topologies such that an app- priate topology is selected, must easily adapt to new semiconductor process nodes, and readily incorporate new topologies. Topology design tools must allow designers to creatively explore new topology ideas as rapidly as possible.

As the miniaturization of semiconductor technology continues, electronic s- tems on chips o?er a more extensive and more complex functionality with better performance, higher frequencies and less power consumption. Whereas digital designers can take full advantage of the availability of design auto- tiontoolstobuildhugesystems,thelackofsupportbycomputerprogramsfor di?erent abstraction levels makes analog design a time-consuming handcraft which limits the possibilities to implement large systems. Various approaches for?ndingoptimalvaluesfortheparametersofanalogcells,likeopamps,have been investigated since the mid-1980s, and they have made their entrance in commercial applications. However, a larger impact on the performance is - pected if tools are developed which operate on a higher abstraction level and consider multiple architectural choices to realize a particular functionality. In this book, the opportunities, conditions, problems, solutions and systematic methodologies for this new generation of analog CAD tools are examined. Theoutlineofthisbookisasfollows. Inthe?rstpart,thecharacteristicsof the analog design process are systematically analyzed and several approaches for automated analog synthesis are summarized. Comparison of their prop- ties with the requirements for high-level synthesis of analog and mixed-signal systems results in a new design paradigm: the high-level design ?ow based on generic behavior. This design approach involves a modeling strategy using generic behavioral models and a synthesis strategy leading to the exploration of a heterogeneous design space containing di?erent architectures. The modeling strategy is further elaborated in Part II.

nalog circuits are fascinating artifacts. They manipulate signals whose informa- Ationcontentisrichcomparedtodigitalsignalsthatcarryminimalamountofinf- mation;theyaredelicateinthatanyperturbationduetoparasiticelements,todelays,to interactionswithotherelementsandwiththeenvironmentmaycauseasigni?cantloss ofinformation. Thedif?cultyindealingwiththeseartifactsistoprotectthemfromall possibleattacks, evenminorones, fromthephysicalworld. Theironyisthattheyare oftenusedtofunnelinformationfromandtothephysicalworldtoandfromtheabstr- tionofthedigitalworldandforthisfunction, theyareirreplaceable. Nowonderthen that analog designers form a club of extraordinary gentlemen where art (or magic?) ratherthanscienceisthesharedtrade. Theyaredif?culttotrainsinceexperienceand intuitionarethetraitsthat characterize them. Andthey have dif?cultiesinexplaining what is the process they use to reach satisfactory results. Tools used for design (s- ulation) are mainly replacing the test benches of an experimental lab. However, the growing complexity of the integrated systems being designed today together with the increasing fragility of analog components brought about by shrinking geometries and reducedpowerconsumptionisposingseverechallengestotraditionalanalogdesigners to produce satisfactory results in a short time. At the same time, the need for expe- enced analog designers has increased constantly since almost all designs, because of integration,docontainanalogcomponents. Thissituationhascreatedastronginterest in developing design methodologies and supporting tools that are based on rigorous, mathematically literate, approaches. Doing so will make it possible to leverage the expertiseofseasonedanalogdesignersandtotrainnewgenerationsfasterandbetter. Inthepast, severalattemptshavebeenmadeinacademia andindustrytocreatethese methodologies and to extend the set of tools available. They have had questionable acceptance in the analog design community. However, recently, a ?urry of start-ups andincreasedinvestmentbyEDAcompaniesinnoveltoolssignalasigni?cantchange inmarketattentiontotheanalogdomain. Ipersonallybelievethattosubstantially- prove quality and design time, tools are simply insuf?cient. A design methodology based on a hierarchy of abstraction layers, successive re?nement between two ad- cent layers, and extensive veri?cation at every layer is necessary. To do so, we need to build theories and models that have strong mathematical foundations. The analog design technology community is as strong as it has ever been.

Systematic Design of Analog IP Blocks introduces a design methodology that can help to bridge the productivity gap. Two different types of designs, depending on the design challenge, have been identified: commodity IP and star IP. Each category requires a different approach to boost design productivity. Commodity IP blocks are well suited to be automated in an analog synthesis environment and provided as soft IP. The design knowledge is usually common knowledge, and reuse is high accounting for the setup time needed for the analog library. Star IP still changes as technology evolves and the design cost can only be reduced by following a systematic design approach supported by point tools to relieve the designer from error-prone, repetitive tasks, allowing him/her to focus on new ideas to push the limits of the design. To validate the presented methodologies, three different industrial-strength applications have been selected and designed accordingly.

It is a great honor to provide a few words of introduction for Dr. Georges Gielen's and Prof. Willy Sansen's book "Symbolic analysis for automated design of analog integrated circuits". The symbolic analysis method presented in this book represents a significant step forward in the area of analog circuit design. As demonstrated in this book, symbolic analysis opens up new possibilities for the development of computer-aided design (CAD) tools that can analyze an analog circuit topology and automatically size the components for a given set of specifications. Symbolic analysis even has the potential to improve the training of young analog circuit designers and to guide more experienced designers through second-order phenomena such as distortion. This book can also serve as an excellent reference for researchers in the analog circuit design area and creators of CAD tools, as it provides a comprehensive overview and comparison of various approaches for analog circuit design automation and an extensive bibliography. The world is essentially analog in nature, hence most electronic systems involve both analog and digital circuitry. As the number of transistors that can be integrated on a single integrated circuit (IC) substrate steadily increases over time, an ever increasing number of systems will be implemented with one, or a few, very complex ICs because of their lower production costs.