Deadlock-Free Message Routing in Multiprocessor Interconnection Networks

@article{Dally1987DeadlockFreeMR,
  title={Deadlock-Free Message Routing in Multiprocessor Interconnection Networks},
  author={William J. Dally and Charles L. Seitz},
  journal={IEEE Transactions on Computers},
  year={1987},
  volume={C-36},
  pages={547-553}
}
  • W. DallyC. Seitz
  • Published 1 May 1987
  • Computer Science
  • IEEE Transactions on Computers
A deadlock-free routing algorithm can be generated for arbitrary interconnection networks using the concept of virtual channels. A necessary and sufficient condition for deadlock-free routing is the absence of cycles in a channel dependency graph. Given an arbitrary network and a routing function, the cycles of the channel dependency graph can be removed by splitting physical channels into groups of virtual channels. This method is used to develop deadlock-free routing algorithms for k-ary n… 

Figures from this paper

Deadlock-free message routing in multicomputer networks

This paper formalizes the concept of message routing and shows a method by which a deadlock-free routing algorithm can be constructed out of a given routing algorithm.

Deadlock-Free Adaptive Routing in Multicomputer Networks Using Virtual Channels

Two deadlock-free adaptive routing algorithms are described that allocate virtual channels using a count of the number of dimension reversals a packet has performed to eliminate cycles in resource dependency graphs and improve virtual channel utilization.

Load balanced deadlock-free deterministic routing of arbitrary networks

An algorithm is presented which will evenly map virtual links onto the routed network so that it will be deadlock-free using the minimum number of buffer classes per physical link, thus maximising the buffer size per virtual link.

Efficient Deadlock-Free Multi-dimensional Interval Routing in Interconnection Networks

The issue of a simultaneous reduction of the compactness and the size is fundamental, worth to investigate and of practical importance, as interval routing and wormhole routing have been realized in INMOS Transputer C104 Router chips.

Generic methodologies for deadlock-free routing

This paper introduces a graph-partitioning generic methodology for developing deadlock-free wormhole routing in an arbitrary network and shows that the methodology can be applied to store-and-forward and virtual cut-through routings as well.

A Sufficient Condition for Deadlock-Free Adaptive Routing in Mesh Networks

It is proved that the proposed flow control algorithm is a sufficient condition for deadlock freedom in any minimal path, adaptive routing algorithms on k-ary, n-cube mesh network.

A Generic Design Methodology for Deadlock-Free Routing in Multicomputer Networks

A generic methodology for defining deadlock-free wormhole routing schemes in any arbitrary network by partitioning a graph into subdigraphs with no cyclic dependencies and selectively assign virtual channels is introduced.

A new deadlock-free fault-tolerant routing algorithm for NoC interconnections

A new deadlock-free fault-tolerant adaptive routing algorithm for the 2D mesh NoC interconnections based on a modified turn model and well known XY algorithm that allows both routing of messages in the networks incorporating the regions not necessarily rectangular, and routing to all nodes which are not completely blocked by faulty nodes.

A Necessary and Sufficient Condition for Deadlock-Free Message Routing in Communication Networks

The deadlock-free routing algorithm is the first to automatically check a necessary and sufficient condition for deadlocked routing, and performs Effective analysis in this network.
...

References

SHOWING 1-10 OF 19 REFERENCES

Deadlock- and livelock-free packet switching networks

This work presents a local controller which is proved to be deadlock- and livelock-free, and guarantees that every packet in the network reaches its destination within a finite amount of time.

Deadlock-free packet switching networks

Several types of deadlock-free controllers are presented, their relationship is studied, and a proof of their optimality with respect to deadlock free controllers using the same set of local parameters is given.

Deadlock Avoidance in Store-and-Forward Networks - I: Store-and-Forward Deadlock

The proposed deadlock avoidance mechanism works for any network topology and any finite routing algorithm and increases the efficiency of buffer use.

A DAG-Based Algorithm for Prevention of Store-and-Forward Deadlock in Packet Networks

It is argued that these costs are imposed infrequently enough and are sufficiently well manageable by heuristic techniques to make this new algorithm an attractive and practical alternative to the older techniques.

Prevention of Deadlocks in Packet-Switched Data Transport Systems

In this paper, necessary and sufficient conditions are given and proven for the use of process ordering and generalized resource ordering techniques to avoid deadlocks in arbitrary systems of interacting processes.

The cube-connected-cycles: A versatile network for parallel computation

This work describes in detail how to program the cube-connected-cycles for efficiently solving a large class of problems, which includes Fast-Fourier-Transform, sorting, permutations, and derived algorithms, and the CCC can also be used as a general purpose parallel processor.

The cube-connected cycles: a versatile network for parallel computation

This work describes in detail how to program the cube-connected cycles for efficiently solving a large class of problems that include Fast Fourier transform, sorting, permutations, and derived algorithms.

On the Performance of k-ary n-cube Interconnection Networks

Low-dimensional k-ary n-cube networks (e.g., tori) have lower latency and higher hot-spot throughput than high-dimensional networks with the same bisection width.

The extension of object-oriented languages to a homogeneous, concurrent architecture

Techniques are developed which permit the execution of general purpose object oriented programs on a homogeneous machine and a distributed, on-the-fly garbage collection algorithm is presented which operates by message passing.