TY - JOUR
T1 - A plane-based broadcast algorithm for multicomputer networks
AU - Al-Dubai, A.
AU - Ould-Khaoua, M.
AU - MacKenzie, L.
PY - 2005/10
Y1 - 2005/10
N2 - Maximising the performance of parallel systems requires matching message-passing algorithms and application characteristics with a suitable underling interconnection network. Broadcast algorithms for wormhole-switched meshes have been widely reported in the literature. However, most of these algorithms handle broadcast in a sequential manner and do not scale well with the network size. As a consequence, many parallel applications cannot be efficiently supported using existing techniques. Motivated by these observations, this paper presents a new efficient broadcast algorithm for the mesh, called the Plane-Based (PB) algorithm. The main feature of this approach is its ability to perform broadcast operation with a high degree of scalability and parallelism. Furthermore, performance is insensitive to the network size, i.e., only three message-passing steps are required to implement a broadcast operation irrespective of the network size. Results from a comparative analysis demonstrate that the PB algorithm exhibits superior performance characteristics over those of the well-known Recursive Doubling and Extending Dominating Node algorithms.
AB - Maximising the performance of parallel systems requires matching message-passing algorithms and application characteristics with a suitable underling interconnection network. Broadcast algorithms for wormhole-switched meshes have been widely reported in the literature. However, most of these algorithms handle broadcast in a sequential manner and do not scale well with the network size. As a consequence, many parallel applications cannot be efficiently supported using existing techniques. Motivated by these observations, this paper presents a new efficient broadcast algorithm for the mesh, called the Plane-Based (PB) algorithm. The main feature of this approach is its ability to perform broadcast operation with a high degree of scalability and parallelism. Furthermore, performance is insensitive to the network size, i.e., only three message-passing steps are required to implement a broadcast operation irrespective of the network size. Results from a comparative analysis demonstrate that the PB algorithm exhibits superior performance characteristics over those of the well-known Recursive Doubling and Extending Dominating Node algorithms.
KW - Collective communication
KW - Mesh
KW - Multicomputer
KW - Performance analysis
KW - Turn model
KW - Wormhole switching
UR - http://www.scopus.com/inward/record.url?scp=25644459290&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=25644459290&partnerID=8YFLogxK
U2 - 10.1016/j.sysarc.2005.01.003
DO - 10.1016/j.sysarc.2005.01.003
M3 - Article
AN - SCOPUS:25644459290
SN - 1383-7621
VL - 51
SP - 617
EP - 632
JO - Journal of Systems Architecture
JF - Journal of Systems Architecture
IS - 10-11
ER -