Publications

Detailed Information

VARIATION AWARE FALSE PATH ANALYSIS IN VLSI DESIGN : 초고집적회로의 지연시간 편차에 따른 허위경로 분석기법

DC Field Value Language
dc.contributor.advisor김태환-
dc.contributor.author정종윤-
dc.date.accessioned2017-07-13T06:53:00Z-
dc.date.available2017-07-13T06:53:00Z-
dc.date.issued2012-08-
dc.identifier.other000000002921-
dc.identifier.urihttps://hdl.handle.net/10371/118844-
dc.description학위논문 (박사)-- 서울대학교 대학원 : 전기·컴퓨터공학부, 2012. 8. 김태환.-
dc.description.abstract반도체 집적회로가 결과를 내기 위해서는 얼마나 많은 시간이 필요한가? 이 물음에 답하기 위해서는 복잡난해한 허위경로 검출문제가 선결되어야 한다. 반도체 공정기술의 발달로 게이트 지연시간 편차가 회로의 동작시간에 큰 영향을 미치게 됨에 따라 허위경로 문제 또한 더욱 복잡해졌다.

Viability 분석기법은 허위경로 문제를 푸는 기법 중 가장 정교한 것이다. 그러나 이 기법은 최장 지연시간 모델에 기초하고 있어 허위경로에 대한 중요한 정보들을 놓칠 우려가 높다. 최근 회로의 지연시간 편차를 고려하기 위해 확률적 지연시간 분석기법 (SSTA)에 대한 연구가 활발히 진행되어왔다. 그러나 이 기법은 회로의 동적 지연시간 변화를 고려할 수 없으므로 회로의 지연시간을 과대평가하게 된다.

본 논문에서는 지연시간 편차에 따른 허위경로 문제에 대해 두 개의 새로운 해법을 제시하고 있다. 그 중 첫 번째는 확률적 viability 분석기법으로써 기존의 viability 분석기법을 획기적으로 개선시킨 것이며, 두 번째는 지연시간의 확률적 동적 지연시간 분석기법(SDTA)으로써 게이트 논리값의 동적 천이의 확률적 변화를 간단한 형식으로 표현가능하게 해주는 기법이다. 이 두 가지 기법은 지연시간 편차에 따른 허위경로 문제에 대해 전혀 다른 두 가지 접근 방식을 제공하고 있다. 이론 및 실험적으로 이 두 기법이 지연시간 편차에 따른 회로의 허위경로들을 효과적으로 검출해낼 수 있음이 입증되었다.
-
dc.description.abstractHow long does an integrated circuit take to produce its result? To answer the question, the difficult and complex false path detection problem must be tackled first. Moreover, the false path problem has become much more complex because the gate delay variation has made a significant impact on the circuit timing, as the technology scales down.

The viability analysis is one of the most sophisticated approaches to the false path detection problem. Since it used the worst-case gate delay model, however, it might miss some important false/true path timing behavior. There has been a lot of research effort on statistical static timing analysis (SSTA) to take into account delay variation in the critical path delay computation. Since SSTA is unable to capture the dynamic timing behavior of a circuit, it is completely blind to false paths, thus, it overestimates the circuit timing.

In this dissertation, two novel approaches to the false path problem under delay variation are proposed. One is the novel statistical viability analysis which is a significantly enhanced version of the viability analysis, and the other is statistical dynamic timing analysis which is able to precisely express the statistical behavior of dynamic transitions at the output of gate into a compact form. Those two approaches provide two different points of view to the false path problem under delay variation. It is theoretically and experimentally shown that those two approaches effectively identify false paths in the major benchmark circuits under delay variation.
-
dc.description.tableofcontentsAbstract i
Contents ii
List of Figures v
List of Tables viii
1 Introduction 1
1.1 False Path Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Requirements for Sensitizing Condition . . . . . . . . . . . . . . . . 1
1.2.1 Correctness . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2.2 Robustness . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.3 Previous Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.4 Contributions of This Dissertation . . . . . . . . . . . . . . . . . . . 3
2 False Path Analysis 5
2.1 False Path Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.2 Sensitizing Conditions . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.2.1 Static Sensitization . . . . . . . . . . . . . . . . . . . . . . . 6
2.2.2 Dynamic Sensitization . . . . . . . . . . . . . . . . . . . . . 7
2.2.3 Viability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
3 False Path Analysis based on Statistical Viability 10
3.1 Viability Analysis under Delay Variation . . . . . . . . . . . . . . . . 10
3.1.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
3.1.2 Viability Function . . . . . . . . . . . . . . . . . . . . . . . 12
3.1.3 Equivalence of AND/OR and MIN/MAX . . . . . . . . . . . 13
3.1.4 Viability Function under Delay Variation . . . . . . . . . . . 14
3.1.5 Statistical Delay Model for Logic gate . . . . . . . . . . . . . 16
3.2 Technique for Fast Viability Analysis . . . . . . . . . . . . . . . . . 16
3.2.1 Step 1: Extracting Good Input Vectors . . . . . . . . . . . . . 17
3.2.2 Step 2: Evaluating Only Viable Late Side Paths . . . . . . . . 18
3.2.3 Complexity Analysis . . . . . . . . . . . . . . . . . . . . . . 22
3.2.4 Example: Statistical Viability Analysis . . . . . . . . . . . . 22
3.3 False Path Aware Statistical Timing Analysis using The Proposed Viability
Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
3.4 Other Useful Applications of The Proposed Viability Analysis . . . . 27
3.4.1 Statistical Gate Sizing . . . . . . . . . . . . . . . . . . . . . 27
3.4.2 Threshold Voltage Assignment . . . . . . . . . . . . . . . . . 28
3.5 Heuristic Choice of the parameter K . . . . . . . . . . . . . . . . . . 29
3.6 Experimental Results . . . . . . . . . . . . . . . . . . . . . . . . . . 29
4 False Path Analysis based on Statistical Dynamic Timing Analysis 38
4.1 Statistical Dynamic Timing Analysis . . . . . . . . . . . . . . . . . . 38
4.1.1 Nondeterministic Timing Behavior under Timing Variation . . 38
4.1.2 Timing Behavior of Logic Gate under Timing Variation . . . . 41
4.1.3 Statistical Timing Model for Logic gate . . . . . . . . . . . . 44
4.1.4 Propagation Theorems of Transition Sequences for Logic Gate 45
4.1.5 Example: Statistical Propagation of Transition Sequences . . 50
4.2 Techniques for Speeding up SDTA Computation . . . . . . . . . . . . 53
4.2.1 Extraction Technique for Sensitizing Input Transition Vectors 53
4.2.2 Efficient Computation Technique for sel Operator . . . . . . 57
4.3 Experimental Results . . . . . . . . . . . . . . . . . . . . . . . . . . 59
5 Conclusion 69
A Proofs of Lemmas 71
A.1 Proof of Lemma 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
A.2 Proof of Lemma 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
A.3 Proof of Lemma 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
B Proofs of Theorems 73
B.1 Proof of Theorem 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
B.2 Proof of Theorem 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
B.3 Proof of Theorem 4 . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
Abstract in Korean 86
-
dc.formatapplication/pdf-
dc.format.extent765231 bytes-
dc.format.mediumapplication/pdf-
dc.language.isoen-
dc.publisher서울대학교 대학원-
dc.subjecttiming analysis-
dc.subjectviability analysis-
dc.subjectfalse path identification-
dc.subjectprocess variation-
dc.subjectSSTA-
dc.subject.ddc621-
dc.titleVARIATION AWARE FALSE PATH ANALYSIS IN VLSI DESIGN-
dc.title.alternative초고집적회로의 지연시간 편차에 따른 허위경로 분석기법-
dc.typeThesis-
dc.contributor.AlternativeAuthorJongyoon Jung-
dc.description.degreeDoctor-
dc.citation.pagesix, 86-
dc.contributor.affiliation공과대학 전기·컴퓨터공학부-
dc.date.awarded2012-08-
Appears in Collections:
Files in This Item:

Altmetrics

Item View & Download Count

  • mendeley

Items in S-Space are protected by copyright, with all rights reserved, unless otherwise indicated.

Share