• We modify the original compressive sensing formulation such that it can be
applied to irregular gate placements.
• We add new constraints to the optimization problem that directly impose
spatial correlations. With these additional constraints, variation estima-
tions improve considerably.
• The proposed post-silicon variation characterization method is fast, inex-
pensive, and non-invasive. It enables a range of new applications. We
introduce a number of novel applications for the proposed method.
The thesis is organized as follows. In Chapter 2, we discuss related work and
preliminaries that are used in the thesis. Preliminaries include the variations
model and the compressive sensing theory. Chapters 3 and 4 introduce our vari-
ation estimation method in power and delay frameworks, respectively. Next, we
discuss a number of applications for the proposed post-silicon variations charac-
terization method in Chapter 5. The evaluation results are presented in Chapter
6. We finally summarize the thesis in Chapter 7.