The turn model for adaptive routing
Journal of the ACM (JACM)
A necessary and sufficient condition for deadlock-free wormhole routing
Journal of Parallel and Distributed Computing
Compressionless Routing: A Framework for Adaptive and Fault-Tolerant Routing
IEEE Transactions on Parallel and Distributed Systems
Characterization of Deadlocks in k-ary n-Cube Networks
IEEE Transactions on Parallel and Distributed Systems
Flexible and Efficient Routing Based on Progressive Deadlock Recovery
IEEE Transactions on Computers
The Odd-Even Turn Model for Adaptive Routing
IEEE Transactions on Parallel and Distributed Systems
A General Theory for Deadlock-Free Adaptive Routing Using a Mixed Set of Resources
IEEE Transactions on Parallel and Distributed Systems
A Cost-Effective Approach to Deadlock Handling in Wormhole Networks
IEEE Transactions on Parallel and Distributed Systems
A generic design methodology for deadlock-free routing in multicomputer networks
Journal of Parallel and Distributed Computing
A Necessary and Sufficient Condition for Deadlock-Free Adaptive Routing in Wormhole Networks
IEEE Transactions on Parallel and Distributed Systems
IEEE Transactions on Parallel and Distributed Systems
A survey and comparison of wormhole routing techniques in a mesh networks
IEEE Network: The Magazine of Global Internetworking
Hi-index | 0.00 |
Wormhole routing has been popular in massively parallel computing systems due to its low packet latency. However, it is subject to deadlock, where packets are waiting for resources in a cyclic form indefinitely. Current deadlock detection techniques are basically dependent on the time-out strategy, thus yielding unignorable number of false deadlock detections especially in heavy network loads or with long packets. Moreover, several packets in a deadlock may be marked as deadlocked, which would saturate the resources allocated for recovery. This paper proposes a simple but more accurate deadlock detection scheme which is less dependent on the time-out value. The proposed scheme presumes deadlock only when a cyclic dependency among blocked packets exists. Consequently, the suggested scheme considerably reduces the probability of detecting false deadlocks over previous schemes, thus enabling more efficient deadlock recovery and higher network throughput. Simulation results are provided to demonstrate the efficiency of the proposed scheme.