TY - GEN
T1 - Exploring concentration and channel slicing in on-chip network router
AU - Kumar, Prabhat
AU - Pan, Yan
AU - Kim, John
AU - Memik, Gokhan
AU - Choudhary, Alok Nidhi
PY - 2009
Y1 - 2009
N2 - Sharing on-chip network resources efficiently is critical in the design of a cost-efficient network on-chip (NoC). Concentration has been proposed for on-chip networks but the trade-off in concentration implementation and performance has not been well understood. In this paper, we describe cost-efficient implementations of concentration and show how external concentration provides a significant reduction in complexity (47% and 36% reduction in area and energy, respectively) compared to previous assumed integrated (high-radix) concentration while degrading overall performance by only 10%. Hybrid implementations of concentration is also presented which provide additional tradeoff between complexity and performance. To further reduce the cost of NoC, we describe how channel slicing can be used together with concentration. We propose virtual concentration which further reduces the complexity - saving area and energy by 69% and 32% compared to baseline mesh and 88% and 35% over baseline concentrated mesh.
AB - Sharing on-chip network resources efficiently is critical in the design of a cost-efficient network on-chip (NoC). Concentration has been proposed for on-chip networks but the trade-off in concentration implementation and performance has not been well understood. In this paper, we describe cost-efficient implementations of concentration and show how external concentration provides a significant reduction in complexity (47% and 36% reduction in area and energy, respectively) compared to previous assumed integrated (high-radix) concentration while degrading overall performance by only 10%. Hybrid implementations of concentration is also presented which provide additional tradeoff between complexity and performance. To further reduce the cost of NoC, we describe how channel slicing can be used together with concentration. We propose virtual concentration which further reduces the complexity - saving area and energy by 69% and 32% compared to baseline mesh and 88% and 35% over baseline concentrated mesh.
UR - http://www.scopus.com/inward/record.url?scp=70349974617&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=70349974617&partnerID=8YFLogxK
U2 - 10.1109/NOCS.2009.5071477
DO - 10.1109/NOCS.2009.5071477
M3 - Conference contribution
AN - SCOPUS:70349974617
SN - 9781424441433
T3 - Proceedings - 2009 3rd ACM/IEEE International Symposium on Networks-on-Chip, NoCS 2009
SP - 276
EP - 285
BT - Proceedings - 2009 3rd ACM/IEEE International Symposium on Networks-on-Chip, NoCS 2009
T2 - 2009 3rd ACM/IEEE International Symposium on Networks-on-Chip, NoCS 2009
Y2 - 10 May 2009 through 13 May 2009
ER -