Ebook: Verification by Error Modeling: Using Testing Techniques in Hardware Verification

1. DESIGN FLOW Integrated circuit (IC) complexity is steadily increasing. ICs incorporating hundreds of millions of transistors, mega-bit memories, complicated pipelined structures, etc., are now in high demand. For example, Intel Itanium II processor contains more than 200 million transistors, including a 3 MB third level cache. A billion transistor IC was said to be “imminently doable” by Intel fellow J. Crawford at Microprocessor Forum in October 2002 [40]. Obviously, designing such complex circuits poses real challenges to engineers. Certainly, no relief comes from the competitive marketplace, with increasing demands for a very narrow window of time (time-to-market) in engineering a ready product. Therefore, a systematic and well-structured approach to designing ICs is a must. Although there are no widely adhered standards for a design flow, most companies have their own established practices, which they follow closely for in-house design processes. In general, however, a typical product cycle includes few milestones. An idea for a new product starts usually from an - depth market analysis of customer needs. Once a window of opportunity is found, product requirements are carefully specified. Ideally, these parameters would not change during the design process. In practice, initial phases of preparing a design specification are susceptible to potential errors, as it is very difficult to grasp all the details in a complex design.

---

This book presents the basis for reusing the test vector generation and simulation for the purpose of implementation verification, to result in a significant timesaving. The book brings the results in the direction of merging manufacturing test vector generation and verification. It first discusses error fault models suitable for approaching the verification by testing methods. Then, it elaborates a proposal for an implicit fault model that uses the Arithmetic Transform representation of a circuit and the faults. Presented is the fundamental link between the error size and the test vector size, which allows parametrizable verification by test vectors. Furthermore, the test vector set is sufficient not only for detecting, but also for diagnosing and correcting the design errors.

---

This book presents the basis for reusing the test vector generation and simulation for the purpose of implementation verification, to result in a significant timesaving. The book brings the results in the direction of merging manufacturing test vector generation and verification. It first discusses error fault models suitable for approaching the verification by testing methods. Then, it elaborates a proposal for an implicit fault model that uses the Arithmetic Transform representation of a circuit and the faults. Presented is the fundamental link between the error size and the test vector size, which allows parametrizable verification by test vectors. Furthermore, the test vector set is sufficient not only for detecting, but also for diagnosing and correcting the design errors.
Content:
Front Matter....Pages i-xv
Introduction....Pages 1-17
Boolean Function Representations....Pages 18-50
Don’t Cares and Their Calculation....Pages 51-69
Testing....Pages 71-102
Design Error Models....Pages 103-128
Design Verification by At....Pages 129-146
Identifying Redundant Gate and Wire Replacements....Pages 147-186
Conclusions and Future Work....Pages 187-190
Back Matter....Pages 191-216

---

This book presents the basis for reusing the test vector generation and simulation for the purpose of implementation verification, to result in a significant timesaving. The book brings the results in the direction of merging manufacturing test vector generation and verification. It first discusses error fault models suitable for approaching the verification by testing methods. Then, it elaborates a proposal for an implicit fault model that uses the Arithmetic Transform representation of a circuit and the faults. Presented is the fundamental link between the error size and the test vector size, which allows parametrizable verification by test vectors. Furthermore, the test vector set is sufficient not only for detecting, but also for diagnosing and correcting the design errors.
Content:
Front Matter....Pages i-xv
Introduction....Pages 1-17
Boolean Function Representations....Pages 18-50
Don’t Cares and Their Calculation....Pages 51-69
Testing....Pages 71-102
Design Error Models....Pages 103-128
Design Verification by At....Pages 129-146
Identifying Redundant Gate and Wire Replacements....Pages 147-186
Conclusions and Future Work....Pages 187-190
Back Matter....Pages 191-216
....