• Refine Query
  • Source
  • Publication year
  • to
  • Language
  • 8
  • 2
  • 1
  • Tagged with
  • 25
  • 25
  • 12
  • 10
  • 9
  • 8
  • 6
  • 5
  • 5
  • 5
  • 4
  • 4
  • 4
  • 4
  • 3
  • About
  • The Global ETD Search service is a free service for researchers to find electronic theses and dissertations. This service is provided by the Networked Digital Library of Theses and Dissertations.
    Our metadata is collected from universities around the world. If you manage a university/consortium/country archive and want to be added, details can be found on the NDLTD website.
1

Static Learning for Problems in VLSI Test and Verification

Syal, Manan 01 July 2005 (has links)
Static learning in the form of logic implications captures Boolean relationships between various gates in a circuit. In the past, logic implications have been applied in several areas of electronic design automation (EDA) including: test-pattern-generation, logic and fault simulation, fault diagnosis, logic optimization, etc. While logic implications have assisted in solving several EDA problems, their usefulness has not been fully explored. We believe that logic implications have not been carefully analyzed in the past, and this lack of thorough investigation has limited their applicability in solving hard EDA problems. In this dissertation, we offer deeper insights into the Boolean relationships exhibited in a circuit, and present techniques to extract their full potential in solving two hard problems in test and verification: (1) Efficient identification of sequentially untestable stuck-at faults, and (2) Equivalence checking of sequential circuits. Additionally, for the dissertation, we define a new concept called multi-cycle path delay faults (M-pdf) for latch based designs with multiple clock domains, and propose an implications-based methodology for the identification of untestable M-pdfs for such designs. One of the main bottlenecks in the efficiency of test-pattern-generation (TPG) is the presence of untestable faults in a design. State-of-the-art automatic test pattern generators (ATPG) spend a lot of effort (in both time and memory) targeting untestable faults before aborting on such faults, or, eventually identifying these faults as untestable (if given enough computational resources). In either case, TPG is considerably slowed down by the presence of untestable faults. Thus, efficient methods to identify untestable faults are desired. In this dissertation, we discuss a number of solutions that we have developed for the purpose of untestable fault identification. The techniques that we propose are fault-independent and explore properties associated with logic implications to derive conclusions about untestable faults. Experimental results for benchmark circuits show that our techniques achieve a significant increase in the number of untestable faults identified, at low memory and computational overhead. The second related problem that we address in this proposal is that of determining the equivalence of sequential circuits. During the design phase, hardware goes through several stages of optimizations (for area, speed, power, etc). Determining the functional correctness of the design after each optimization step by means of exhaustive simulation can be prohibitively expensive. An alternative to prove functional correctness of the optimized design is to determine the design's functional equivalence w.r.t. some golden model which is known to be functionally correct. Efficient techniques to perform this process, known as equivalence checking, have been investigated in the research community. However, equivalence checking of sequential circuits still remains a challenging problem. In an attempt to solve this problem, we propose a Boolean SAT (satisfiability) based framework that utilizes logic implications for the purpose of sequential equivalence checking. Finally, we define a new concept called multi-cycle path-delay faults (M-pdfs). Traditionally, path delay faults have been analyzed for flip-flop based designs over the boundary of a single clock cycle. However, path delay faults may span multiple clock cycles, and a technique is desired to model and analyze such path delay faults. This is especially essential for latch based designs with multiple clock domains, because the problem of identifying untestable faults is more complex in such design environments. In this dissertation, we propose a three-step methodology to identify untestable M-pdfs in latch-based designs with multiple clocks using logic implications. / Ph. D.
2

ATPG based Preimage Computation: Efficient Search Space Pruning using ZBDD

Chandrasekar, Kameshwar 06 August 2003 (has links)
Preimage Computation is a fundamental step in Formal Verification of VLSI designs. Conventional OBDD-based methods for Formal Verification suffer from spatial explosion, since large designs can blow up in terms of memory. On the other hand, SAT/ATPG based methods are less demanding on memory. But the run-time can be huge for these methods, since they must explore an exponential search space. In order to reduce this temporal explosion of SAT/ATPG based methods, efficient learning techniques are needed. Conventional ATPG aims at computing a single solution for its objective. In preimage computation, we must enumerate all solutions for the target state during the search. Similar sub-problems often occur during preimage computation that can be identified by the internal state of the circuit. Therefore, it is highly desirable to learn from these search-states and avoid repeated search of identical solution/conflict subspaces, for better performance. In this thesis, we present a new ZBDD based method to compactly store and efficiently search previously explored search-states. We learn from these search-states and avoid repeating subsets and supersets of previously encountered search spaces. Both solution and conflict subspaces are pruned based on simple set operations using ZBDDs. We integrate our techniques into a PODEM based ATPG engine and demonstrate their efficiency on ISCAS '89 benchmark circuits. Experimental results show that upto 90% of the search-space is pruned due to the proposed techniques and we are able to compute preimages for target states where a state-of-the-art technique fails. / Master of Science
3

Behavioral Model Equivalence Checking for Large Analog Mixed Signal Systems

Singh, Amandeep 2011 May 1900 (has links)
This thesis proposes a systematic, hierarchical, optimization based semi-formal equivalence checking methodology for large analog/mixed signal systems such as phase locked loops (PLL), analog to digital convertors (ADC) and input/output (I/O) circuits. I propose to verify the equivalence between a behavioral model and its electrical implementation over a limited, but highly likely, input space defined as the Constrained Behavioral Input Space. Furthermore, I clearly distinguish between the behavioral and electrical domains and define mapping functions between the two domains to allow for calculation of deviation between the behavioral and electrical implementation. The verification problem is then formulated as an optimization problem which is solved by interfacing a sequential quadratic programming (SQP) based optimizer with commercial circuit simulation tools, such as CADENCE SPECTRE. The proposed methodology is then applied for equivalence checking of a PLL as a test case and results are shown which prove the correctness of the proposed methodology.
4

Formal Verification Methodologies for NULL Convention Logic Circuits

Le, Son Ngoc January 2020 (has links)
NULL Convention Logic (NCL) is a Quasi-Delay Insensitive (QDI) asynchronous design paradigm that aims to tackle some of the major problems synchronous designs are facing as the industry trend of increased clock rates and decreased feature size continues. The clock in synchronous designs is becoming increasingly difficult to manage and causing more power consumption than ever before. NCL circuits address some of these issues by requiring less power, producing less noise and electro-magnetic interference, and being more robust to Process, Voltage, and Temperature (PVT) variations. With the increase in popularity of asynchronous designs, a formal verification methodology is crucial for ensuring these circuits operate correctly. Four automated formal verification methodologies have been developed, three to ensure delay-insensitivity of an NCL circuit (i.e., prove Input-Completeness, Observability, and Completion-Completeness properties), and one to aid in proving functional equivalence between an NCL circuit and its synchronous counterpart. Note that an NCL circuit can be functionally correct and still not be input-complete, observable, or completion-complete, which could cause the circuit to operate correctly under normal conditions, but malfunction when circuit timing drastically changes (e.g., significantly reduced supply voltage, extreme temperatures). Since NCL circuits are implemented using dual-rail logic (i.e., 2 wires, rail0 and rail1, represent one bit of data), part of the functional equivalence verification involves ensuring that the NCL rail0 logic is the inverse of its rail1 logic. Equivalence verification optimizations and alternative invariant checking methods were investigated and proved to decrease verification times of identical circuits substantially. This work will be a major step toward NCL circuits being utilized more frequently in industry, since it provides an automated verification method to prove correctness of an NCL implementation and equivalence to its synchronous specification, which is the industry standard.
5

Strategies for SAT-Based Formal Verification

Vimjam, Vishnu Chaithanya 13 February 2007 (has links)
Verification of digital hardware designs is becoming an increasingly complex task as the designs are incorporating more functionality, becoming complex and growing larger in size. Today, verification remains a bottleneck in meeting time-to-market requirements and consumes more than 70% of the overall design-costs. Traditionally, verification has been done using simulation-based approaches, where a set of appropriate test-stimuli is used by the designer. As the designs become more complex, however, simulation-based techniques often fail to capture corner-case errors. Furthermore, unless exhaustively tested, these approaches do not guarantee the correctness of a system with respect to its specifications. As a consequence, formal methods for design verification have been sought after. In formal verification, the conformance of a design to a given set of specifications is proven mathematically, thereby leaving no room for unexplored search spaces. Despite the exponential time/memory complexities often involved within the formal approaches, they have shown promise in capturing subtle bugs, which were missed otherwise. In this dissertation, we focus on Boolean Satisfiability (SAT) based formal verification, which has gained tremendous importance in the recent past. Importantly, SAT-based approaches often alleviate the memory explosion problem, which had been a bottleneck of the traditional symbolic (Binary Decision Diagram based) approaches. In SAT-based techniques, the set of verification tasks are converted into a set of Boolean formulae, which are checked for satisfiability using a SAT solver. These problems are often NP-complete and are prone to an explosion in the required run-time. To overcome this, we propose novel strategies which utilize both structural and logical information of a sequential circuit. In particular, we devise techniques to extract non-trivial invariants of a design, strengthen properties such that they can be proven faster and interleave bounded reachability analysis with bounded model checking. We provide the necessary algorithms and implementation details in order to automate the proposed techniques. Experiments conducted on a variety of benchmark circuits show that orders of magnitude improvement in overall run-times can be achieved via our techniques compared to the existing state-of-the-art SAT-based approaches. / Ph. D.
6

Novel RTD-Based Threshold Logic Design and Verification

Zheng, Yexin 06 May 2008 (has links)
Innovative nano-scale devices have been developed to enhance future circuit design to overcome physical barriers hindering complementary metal-oxide semiconductor (CMOS) technology. Among the emerging nanodevices, resonant tunneling diodes (RTDs) have demonstrated promising electronic features due to their high speed switching capability and functional versatility. Great circuit functionality can be achieved through integrating heterostructure field-effect transistors (HFETs) in conjunction with RTDs to modulate effective negative differential resistance (NDR). However, RTDs are intrinsically suitable for implementing threshold logic rather than Boolean logic which has dominated CMOS technology in the past. To fully take advantage of such emerging nanotechnology, efficient design methodologies and design automation tools for threshold logic therefore become essential. In this thesis, we first propose novel programmable logic elements (PLEs) implemented in threshold gates (TGs) and multi-threshold threshold gates (MTTGs) by exploring RTD/ HFET monostable-bistable transition logic element (MOBILE) principles. Our three-input PLE can be configured through five control bits to realize all the three-variable logic functions, which is, to the best of our knowledge, the first single RTD-based structure that provides complete logic implementation. It is also a more efficient reconfigurable circuit element than a general look-up table which requires eight configuration bits for three-variable functions. We further extend the design concept to construct a more versatile four-input PLE. A comprehensive comparison of three- and four-input PLEs provides an insightful view of design tradeoffs between performance and area. We present the mathematical proof of PLE's logic completeness based on Shannon Expansion, as well as the HSPICE simulation results of the programmable and primitive RTD/HFET gates that we have designed. An efficient control bit generating algorithm is developed by using a special encoding scheme to implement any given logic function. In addition, we propose novel techniques of formulating a given threshold logic in conjunctive normal form (CNF) that facilitates efficient SAT-based equivalence checking for threshold logic networks. Three different strategies of CNF generation from threshold logic representations are implemented. Experimental results based on MCNC benchmarks are presented as a complete comparison. Our hybrid algorithm, which takes into account input symmetry as well as input weight order of threshold gates, can efficiently generate CNF formulas in terms of both SAT solving time and CNF generating time. / Master of Science
7

Formal Verification Techniques for Reversible Circuits

Limaye, Chinmay Avinash 27 June 2011 (has links)
As the number of transistors per unit chip area increases, the power dissipation of the chip becomes a bottleneck. New nano-technology materials have been proposed as viable alternatives to CMOS to tackle area and power issues. The power consumption can be minimized by the use of reversible logic instead of conventional combinational circuits. Theoretically, reversible circuits do not consume any power (or consume minimal power) when performing computations. This is achieved by avoiding information loss across the circuit. However, use of reversible circuits to implement digital logic requires development of new Electronic Design Automation techniques. Several approaches have been proposed and each method has its own pros and cons. This often results in multiple designs for the same function. Consequently, this demands research in efficient equivalence checking techniques for reversible circuits. This thesis explores the optimization and equivalence checking of reversible circuits. Most of the existing synthesis techniques work in two steps — generate an original, often sub-optimal, implementation for the circuit followed optimization of this design. This work proposes the use of Binary Decision Diagrams for optimization of reversible circuits. The proposed technique identifies repeated gate (trivial) as well as non-contiguous redundancies in a reversible circuit. Construction of a BDD for a sub-circuit (obtained by sliding a window of fixed size over the circuit) identifies redundant gates based upon the redundant variables in the BDD. This method was unsuccessful in identifying any additional redundancies in benchmark circuits; however, hidden non-contiguous redundancies were consistently identified for a family of randomly generated reversible circuits. As of now, several research groups focus upon efficient synthesis of reversible circuits. However, little work has been done in identification of redundant gates in existing designs and the proposed peephole optimization method stands among the few known techniques. This method fails to identify redundancies in a few cases indicating the complexity of the problem and the need for further research in this area. Even for simple logical functions, multiple circuit representations exist which exhibit a large variation in the total number of gates and circuit structure. It may be advantageous to have multiple implementations to provide flexibility in choice of implementation process but it is necessary to validate the functional equivalence of each such design. Equivalence checking for reversible circuits has been researched to some extent and a few pre-processing techniques have been proposed prior to this work. One such technique involves the use of Reversible Miter circuits followed by SAT-solvers to ascertain equivalence. The second half of this work focuses upon the application of the proposed reduction technique to Reversible Miter circuits as a pre-processing step to improve the efficiency of the subsequent SAT-based equivalence checking. / Master of Science
8

Game semantics based equivalence checking of higher-order programs

Hopkins, David G. B. January 2012 (has links)
This thesis examines the use of game semantics for the automatic equivalence checking of higher-order programs. Game semantics has proved to be a powerful method for constructing fully abstract models of logics and programming languages. Furthermore, the concrete nature of the semantics lends itself to algorithmic analysis. The game-semantic model can be used to identify fragments of languages which have a decidable observational equivalence problem. We investigate decidability results for different languages as well as the efficiency of these algorithms in practice. First we consider the call-by-value higher-order language with state, RML. This can be viewed as a canonical restriction of Standard ML to ground-type references. The O-strict fragment of RML is the largest set of type sequents for which, in the game-semantic denotation, justification pointers from O-moves are always uniquely reconstructible from the underlying move sequence. The O-strict fragment is surprisingly expressive, including higher-order types and difficult examples from the literature. By representing strategies as Visibly Pushdown Automata (VPA) we show that observational equivalence of O-strict terms is decidable (and in fact is ExpTime-complete). We then consider extensions of the O-strict fragment. Adding general recursion or using most non-O-strict types leads to undecidability. However, a limited form of recursion can be added while still preserving decidability (although the full power of DPDA is required). Next we examine languages with non-local control. This involves adding call/cc to our language and is known to correspond to dropping the game-semantic bracketing condition. In the call-by-name game-semantic model of Idealized Algol (IA), in which answers cannot justify questions, the visibility condition still implies a form of weak bracketing. By making bracketing violations explicit we show that we can still model the entire third-order fragment using VPA. We have also implemented tools based on these algorithms. Our model checkers Homer and Hector perform equivalence checking for third-order IA and O-strict RML respectively. Homer uses a naive explicit state method whereas Hector takes advantage of on-the-fly model checking. Our tools perform well on small yet challenging examples. On negative instances, the on-the-fly approach allows Hector to outperform Homer. To improve their performance, we also consider using ideas from symbolic execution. We propose a representation for finite automata using transitions labelled with formulas and guards which aims to take advantage of the symmetries of the game-semantic model so that strategies can be represented compactly. We refer to this representation as Symbolically Executed Automata (SEA). Using SEA allows much larger data types to be handled but is not as effective on larger examples with small data types.
9

Dominator-based Algorithms in Logic Synthesis and Verification

Krenz-Bååth, René January 2007 (has links)
Today's EDA (Electronic Design Automation) industry faces enormous challenges. Their primary cause is the tremendous increase of the complexity of modern digital designs. Graph algorithms are widely applied to solve various EDA problems. In particular, graph dominators, which provide information about the origin and the end of reconverging paths in a circuit graph, proved to be useful in various CAD (Computer Aided Design) applications such as equivalence checking, ATPG, technology mapping, and power optimization. This thesis provides a study on graph dominators in logic synthesis and verification. The thesis contributes a set of algorithms for computing dominators in circuit graphs. An algorithm is proposed for finding absolute dominators in circuit graphs. The achieved speedup of three orders of magnitude on several designs enables the computation of absolute dominators in large industrial designs in a few seconds. Moreover, the computation of single-vertex dominators in large multiple-output circuit graphs is considerably improved. The proposed algorithm reduces the overall runtime by efficiently recognizing and re-using isomorphic structures in dominator trees rooted at different outputs of the circuit graph. Finally, common multiple-vertex dominators are introduced. The algorithm to compute them is faster and finds more multiple-vertex dominators than previous approaches. The thesis also proposes new dominator-based algorithms in the area of decomposition and combinational equivalence checking. A structural decomposition technique is introduced, which finds all simple-disjoint decompositions of a Boolean function which are reflected in the circuit graph. The experimental results demonstrate that the proposed technique outperforms state-of-the-art functional decomposition techniques. Finally, an approach to check the equivalence of two Boolean functions probabilistically is investigated. The proposed algorithm partitions the equivalence check employing dominators in the circuit graph. The experimental results confirm that, in comparison to traditional BDD-based equivalence checking methods, the memory consumption is considerably reduced by using the proposed technique. / QC 20100804
10

Graph dominators in logic synthesis and verification

Krenz, René January 2004 (has links)
<p>This work focuses on the usage of dominators in circuit graphs in order to reduce the complexity of synthesis and verification tasks. One of the contributions of this thesis is a new algorithm for computing multiple-vertex dominators in circuit graphs. Previous algorithms, based on single-vertex dominators suffer from their rare appearance in many circuits. The presented approach searches efficiently for multiple-vertex dominators in circuit graphs. It finds dominator relations, where algorithms for computing single-vertex dominators fail. Another contribution of this thesis is the application of dominators for combinational equivalence checking based on the arithmetic transform. Previous algorithms rely on representations providing an explicit or implicit disjoint function cover, which is usually excessive in memory requirements. The new algorithm allows a partitioned evaluation of the arithmetic transform directly on the circuit graph using dominator relations. The results show that the algorithm brings significant improvements in memory consumption for many benchmarks. Proper cuts are used in many areas of VLSI. They provide cut points, where a given problem can be split into two disjoint sub-problems. The algorithm proposed in this thesis efficiently detects proper cuts in a circuit graph and is based on a novel concept of a reduced dominator tree. The runtime of the algorithm is less than 0.4 seconds for the largest benchmark circuit. The final contribution of this thesis is the application of the proper cut algorithm as a structural method to decompose a Boolean function, represented by a circuit graph. In combination with a functional approach, it outperforms previous methods, which rely on functional decomposition only.</p>

Page generated in 0.1258 seconds