Name
Abstract | ||
|---|---|---|
We propose a framework that allows dual-layer cooperative error control in a nanoscale network-on-chip (NoC), to simultaneously improve reliability, performance and energy efficiency. The framework combines a configurable routing algorithm in the network layer and an adaptive error control coding scheme in the datalink layer. When the network layer detects an increase in network congestion, either the routing algorithm is switched from deterministic to partially adaptive routing or the error controls schemes is adjusted to a powerful mode. In high noise conditions, the adaptive error control in the datalink layer is switched to a more powerful error control mode, cooperating with the configurable routing algorithm. The resulting dual-layer framework allows reliability and network congestion to be managed dynamically. Simulation results show that the proposed method improves the average latency up to 24% and the average energy up to 27% compared to the single layer error control, with 8% area overhead. |
Year | DOI | Venue |
|---|---|---|
2009 | 10.1109/DFT.2009.48 | DFT |
Keywords | Field | DocType |
network congestion,adaptive error control,integrated circuit reliability,average latency,reliable nanoscale networks-on-chip,routing algorithm,datalink layer,single layer error control,average energy,configurable routing algorithm,adaptive routing,noise conditions,powerful error control mode,vlsi,integrated circuit noise,dual-layer cooperative error control,adaptive error control coding scheme,network layer,energy efficiency,network-on-chip,nanoscale networks-on-chip reliability,encoding,noise,error correction,dynamic simulation,network on chip,error control,routing,energy efficient | Computer science,Efficient energy use,Data link layer,Network layer,Network on a chip,Real-time computing,Electronic engineering,Error detection and correction,Network congestion,Very-large-scale integration,Encoding (memory) | Conference |
ISSN | ISBN | Citations |
1550-5774 | 978-0-7695-3839-6 | 0 |
PageRank | References | Authors |
0.34 | 15 | 2 |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Qiaoyan Yu | 1 | 174 | 28.58 |
| Paul Ampadu | 2 | 285 | 28.55 |