2:00 PM on Tuesday, December 1, 2009
Location: 330 Durham

Master of Science Thesis Defense for Sandesh Prabhakar

Committee Members:
Michael Hsiao, Chair
Patrick Schaumont, Member
Christopher Wyatt, Member

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. In this thesis, we propose 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. We also propose a new multiplexer-based trace signal interconnection scheme and a new heuristic for trace signal selection based on implication-based correlation. As a result, we can effectively trace twice as many signals with the same trace buffer width. We then propose a SAT-based greedy heuristic to prune the selected trace signal list further to take into account those multi-node implications. A state restoration algorithm is developed for the multiplexer-based trace signal interconnection scheme. Experimental result showed that the proposed approaches select the trace signals effectively, giving a high restoration percentage compared with other techniques. We finally propose a lossless compression technique to increase the capacity of the trace buffer. Source transformation techniques that improve the existing FDR code based compression technique are used. Source transformation reduces the entropy of the data to be compressed and hence, improves the compression ratio. We implement the method on hardware and observe that the area overhead of the compressor is less compared to dictionary-based compression techniques like LZ. Experimental result showed that this improved method gives a better compression percentage than existing FDR code based technique and GZIP for compression of the traced signal data.

Contact: hsiao(at)vt( dot )edu