PerFT-N: Low-overhead Permanent Fault-Tolerance Mechanism for Neural Processing Units.
Haojie JianChao ChenZheng WangPengfei WuPublished in: ACM Great Lakes Symposium on VLSI (2024)
Keyphrases
- fault tolerance
- low overhead
- load balancing
- processing units
- fault tolerant
- distributed systems
- peer to peer
- parallel computing
- distributed computing
- parallel processing
- mobile agents
- shared memory
- computing systems
- response time
- high reliability
- communication cost
- single point of failure
- grid computing
- multiple types
- high performance computing
- low cost
- mobile devices
- artificial intelligence
- real time