Can't See the Forest for Trees: State Restoration's Limitations in Post-silicon Trace Signal Selection

S. Ma, D. Pal, R. Jiang, S. Ray, and S. Vasudevan

In D. Marculescu, F. Liu, and S. Parameswaran editors, 34th International Conference on Computer-Aided Design (ICCAD 2015), Austin, TX, USA, November 2015, pages 1-8. IEEE.

© 2015 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE.


Abstract

Restoration Ratio (SRR) has been the de facto standard for evaluating quality of signals selected for post-silicon tracing and debug. Given a set S of selected signals, SRR measures the fraction of (gate-level) design states that can be inferred from observing signals in S at each cycle. Unfortunately, in spite of its widespread use, we found that SRR is intrinsically unsuitable as a metric for evaluating trace signal quality, as it captures neither the higher-level functionality of the design nor the constraints and requirements on trace signals imposed by architectural, physical, or security requirements. In this paper, we argue with strong empirical evidence that SRR must be replaced by a metric that closely models high-level behavioral coverage. We propose assertion coverage as a first step in this direction. We also present a new algorithm, based on Pagerank, for post-silicon trace selection. Pagerank is not designed to maximize SRR. We found that Pagerank has upto 70% higher behavioral coverage than SRR optimizing methods, and the RTL PageRank has upto 30% higher behavioral coverage than the netlist PageRank algorithm. Assertion coverage of PageRank RTL is upto 50% while SRR based methods have less than 5% assertion coverage.

Relevant files