Performance of multicomputer networks under pin-out constraints

Abraham, Santosh G.; Padmanabhan, Krishnan

doi:10.1016/0743-7315(91)90128-v

Cited by 62 publications

(57 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Other researchers, including Abraham [2] and Agrawal [3], have conducted similar studies to Dally's and arrived at the same conclusion. However, they have also argued that while the wiring density constraint is certainly applicable where an entire network is implemented on a single VLSI-chip, this is not the case in the currently more realistic situation where a network has to be partitioned over many chips.…”

Section: Performance Comparisonsupporting

confidence: 59%

A comparative performance analysis of n-cubes and star graphs

Kiasari

Sarbazi-Azad

2006

Proceedings 20th IEEE International Parallel &Amp; Distributed Processing Symposium

View full text Add to dashboard Cite

show abstract

Section: Performance Comparisonsupporting

confidence: 59%

A comparative performance analysis of n-cubes and star graphs

Kiasari

Sarbazi-Azad

2006

Proceedings 20th IEEE International Parallel &Amp; Distributed Processing Symposium

View full text Add to dashboard Cite

show abstract

“…• A medium size of N = 512 nodes; configured as a 23 × 23 torus ‡ , 8 × 8 × 8 torus and 2 9 hypercube.…”

Section: Resultsmentioning

confidence: 99%

“…Most existing studies [9,10,12,14,17] have assumed the uniform traffic pattern. This is unrealistic in practice as there are many real-world parallel applications that exhibit non-uniform traffic patterns, which often produce hotspots in the network [22].…”

mentioning

confidence: 99%

Towards a more realistic comparative analysis of multicomputer networks

Sarbazi-Azad

Ould‐Khaoua

Mackenzie

2004

Concurrency and Computation

View full text Add to dashboard Cite

SUMMARYSeveral studies have examined the relative performance merits of the torus and hypercube taking into account the channel bandwidth constraints imposed by implementation technology. While the torus has been shown to outperform the hypercube under the constant wiring density constraint, the opposite conclusion has been reached when the constant pin-out constraint is considered. However, these studies have assumed a pure uniform traffic pattern and deterministic routing. The 'uniform traffic' assumption is not always justifiable in practice as there are many real-world parallel applications that exhibit nonuniform traffic patterns, which can create unbalanced traffic such as hotspots in the network. This paper re-examines the performance merits of the torus and hypercube in the presence of hotspot traffic. The comparative analysis is based on fully adaptive routing as this has been gaining popularity in recent practical multicomputers. Moreover, it uses a new cost model that takes into account the implementation cost of the network and its routers. The results reveal that for moderate and large system sizes, lower dimensional k-ary n-cubes (e.g. 2D torus) always outperform their higher dimensional counterparts even under the pin-out constraint.

show abstract

“…However, technology that allows an entire system to be implemented on a single chip will not be achievable for many years. The authors in [1,2] have identified pin-out as the constraint applicable to current multi-chip technology, and they have reported that when nodes are subjected to this condition it is the hypercube that exhibits the superior performance.…”

Section: Introductionmentioning

confidence: 97%

Distributed Crossbar Switch Hypermeshes: Efficient Networks for Large-Scale Multicomputers

Ould‐Khaoua¹

2001

Journal of Parallel and Distributed Computing

View full text Add to dashboard Cite

Performance of multicomputer networks under pin-out constraints

Cited by 62 publications

References 7 publications

A comparative performance analysis of n-cubes and star graphs

A comparative performance analysis of n-cubes and star graphs

Towards a more realistic comparative analysis of multicomputer networks

Distributed Crossbar Switch Hypermeshes: Efficient Networks for Large-Scale Multicomputers

Contact Info

Product

Resources

About