AWAIT: An Ultra-Lightweight Soft-Error Mitigation Mechanism for Network-on-Chip Links
Autor: | Rene Pihlak, Siavoosh Payandeh Azad, Gert Jervan, Karl Janson, Jaan Raik, Behrad Niazmand |
---|---|
Jazyk: | angličtina |
Předmět: |
010302 applied physics
Computer science Distributed computing Fault tolerance Hardware_PERFORMANCEANDRELIABILITY 02 engineering and technology Fault (power engineering) 01 natural sciences Fault detection and isolation 020202 computer hardware & architecture Network on a chip Soft error 0103 physical sciences Scalability 0202 electrical engineering electronic engineering information engineering Overhead (computing) Transient (oscillation) |
Zdroj: | 2018 13th International Symposium on Reconfigurable Communication-centric Systems-on-Chip (ReCoSoC) ReCoSoC |
DOI: | 10.1109/recosoc.2018.8449374 |
Popis: | Networks-on-Chip have become a widely accepted communication paradigm for many-core Systems-on-Chip. However, with the ever-shrinking transistor size, the network's sensitivity to transient faults on the physical links cannot be ignored since even a single transient fault can lead to a network-wide congestion and a system failure. This paper proposes the AWAIT mechanism, an ultra-lightweight transient fault mitigation mechanism for Network-on-Chip links. The proposed mechanism covers all single event transients. The experimental results show that the AWAIT mechanism prevents network-wide failure even in harsh environments (up to 80 million random faults on links per second). The AWAIT mechanism is also scalable and imposes only 5.1 % area overhead with very negligible critical path delay overhead. |
Databáze: | OpenAIRE |
Externí odkaz: |