Using Non-trivial Logic Implications for Trace Buffer-Based Silicon Debug

Abstract

An effective silicon debug technique uses a trace buffer to monitor and capture a portion of the circuit response during its functional, post-silicon operation. Due to the limited space of the available trace buffer, selection of the critical trace signals plays an important role in both minimizing the number of signals traced and maximizing the observability/restorability of other untraced signals during post-silicon validation. This paper presents a new method for trace buffer signal selection for the purpose of post-silicon debug. The selection is performed by favoring those signals with the most number of implications that are not implied by other signals. Then, based on the values of the traced signals during silicon debug, we introduce an algorithm which uses a SAT-based multi-node implication engine to restore the values of untraced signals across multiple time-frames.Experimental results for sequential benchmark circuits showed that the proposed approach selects the trace signals effectively, giving a high restoration percentage compared with other techniques.

Simulation-based signal selection for state restoration in silicon debug

“…They also introduced the first trace signal selection algorithm striving to maximize the amount of restored state. Further research in this area has produced several improved solutions for automatic signal selection [8], [9], [10], all sharing the goal of improving the SRR.…”

“…The selection algorithm grows a region of flip-flops in a greedy fashion based on this metric, while an adjustment mechanism accounts for flipflops that are already selected in the region and updates the path's probabilities accordingly. Another solution presented in [9] estimates the visibility of non-traced nodes by non-trivial logic implications of flip-flop values. However, [9] assumes that in addition to trace signals, all primary input values for every cycle are known to the restoration algorithm.…”

“…Another solution presented in [9] estimates the visibility of non-traced nodes by non-trivial logic implications of flip-flop values. However, [9] assumes that in addition to trace signals, all primary input values for every cycle are known to the restoration algorithm. Our proposed solution is fundamentally different from these previous ones as it relies on simulation for estimation instead of a probabilistic metric.…”

“…Most signal selection algorithms presented in the literature so far [14], [8], [9], [10] share a common structure. First, a metric is devised to estimate the state restoration capacity of a given set of signals; second, a greedy selection process guided by the metric is used to converge to a locally-optimal selection.…”

“…This measure has been adopted by subsequent research to compare the quality of other solutions. Further research [8], [9], [10] has proposed several automated trace signal selection methods based on different heuristics for estimating the state restoration capabilities of a group of signals. These research solutions share a common structure: (i) a metric to estimate the state restoration capability of a set of state elements and (ii) the use of the metric in a greedy selection process to evaluate candidate set of signals and converge to a final selection.…”