TY - GEN
T1 - Transaction-level error susceptibility for bus-based system-on-chip
T2 - 2010 International Computer Symposium, ICS 2010
AU - Zheng, Shi Qun
AU - Lin, Ing Chao
PY - 2010
Y1 - 2010
N2 - System-on-Chip architectures have traditionally relied upon bus-based interconnect for their communication needs. The increasing bus frequencies and load on the bus calls for focus on reliability issues in such bus-based systems. As technology advances and transistor geometry shrinks, both single-bit and multi-bit error rate increase significantly. The scant research on mulit-bit errors calls for more attention about them. In this paper, we compare the consequences of a single-bit and multi-bit error and provide a detail analysis of a multi-bit error on the bus system during the course of different transactions. Such transaction based analysis helps us to develop an effective prediction methodology to predict the effect of a multi-bit error on any application running on a bus based architecture. We demonstrate that our transaction based prediction scheme works with an average accuracy of 88% over all the benchmarks when compared with the actual simulation results.
AB - System-on-Chip architectures have traditionally relied upon bus-based interconnect for their communication needs. The increasing bus frequencies and load on the bus calls for focus on reliability issues in such bus-based systems. As technology advances and transistor geometry shrinks, both single-bit and multi-bit error rate increase significantly. The scant research on mulit-bit errors calls for more attention about them. In this paper, we compare the consequences of a single-bit and multi-bit error and provide a detail analysis of a multi-bit error on the bus system during the course of different transactions. Such transaction based analysis helps us to develop an effective prediction methodology to predict the effect of a multi-bit error on any application running on a bus based architecture. We demonstrate that our transaction based prediction scheme works with an average accuracy of 88% over all the benchmarks when compared with the actual simulation results.
UR - http://www.scopus.com/inward/record.url?scp=79851488813&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=79851488813&partnerID=8YFLogxK
U2 - 10.1109/COMPSYM.2010.5685428
DO - 10.1109/COMPSYM.2010.5685428
M3 - Conference contribution
AN - SCOPUS:79851488813
SN - 9781424476404
T3 - ICS 2010 - International Computer Symposium
SP - 670
EP - 675
BT - ICS 2010 - International Computer Symposium
Y2 - 16 December 2010 through 18 December 2010
ER -