Increasing IC densities necessitate diagnosis methodologies with enhanceddefect locating capabilities. Yet the computational effort expended in extracting diagnosticinformation and the stringent storage requirements c...Increasing IC densities necessitate diagnosis methodologies with enhanceddefect locating capabilities. Yet the computational effort expended in extracting diagnosticinformation and the stringent storage requirements constitute major concerns due to the tremendousnumber of faults in typical ICs. In this paper, we propose an RT-level diagnosis methodology capableof responding to these challenges. In the proposed scheme, diagnostic information is computed on agrouped fault effect basis, enhancing both the storage and the computational aspects. The faulteffect grouping criteria are identified based on a module structure analysis, improving thepropagation ability of the diagnostic information through RT modules. Experimental results show thatthe proposed methodology provides superior speed-ups and significant diagnostic informationcompression at no sacrifice in diagnostic resolution, compared to the existing gate-level diagnosisapproaches.展开更多
文摘Increasing IC densities necessitate diagnosis methodologies with enhanceddefect locating capabilities. Yet the computational effort expended in extracting diagnosticinformation and the stringent storage requirements constitute major concerns due to the tremendousnumber of faults in typical ICs. In this paper, we propose an RT-level diagnosis methodology capableof responding to these challenges. In the proposed scheme, diagnostic information is computed on agrouped fault effect basis, enhancing both the storage and the computational aspects. The faulteffect grouping criteria are identified based on a module structure analysis, improving thepropagation ability of the diagnostic information through RT modules. Experimental results show thatthe proposed methodology provides superior speed-ups and significant diagnostic informationcompression at no sacrifice in diagnostic resolution, compared to the existing gate-level diagnosisapproaches.
