.png){kind=icon}
.png){kind=icon}
Return
Fast Error Propagation Probability Estimates by Answer Set Programming and Approximate Model Counting
.png){kind=icon}
OA
AI
DOI：10.1109/ACCESS.2022.3174564.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
We present a method employing Answer Set Programming in combination with Approximate Model Counting for fast and accurate calculation of error propagation probabilities in digital circuits. By an efficient problem encoding, we achieve an input data format similar to a Verilog netlist so that extensive preprocessing is avoided. By a tight interconnection of our application with the underlying solver, we avoid iterating over fault sites and reduce calls to the solver. Several circuits were analyzed with varying numbers of considered cycles and different degrees of approximation. Our experiments show, that the runtime can be reduced by approximation by a factor of 91, whereas the error compared to the exact result is below 1%.
Keywords:
Circuit faults
Integrated circuit modeling
Programming
Analytical models
Search problems
Flip-flops
Encoding
Answer set programming
approximate model counting
error propagation
radhard design
reliability analysis
selective fault tolerance
single event upsets