Realizing Common Communication Patterns in Partitioned Optical Passive Stars (POPS) Networks
IEEE Transactions on Computers
Matrix Multiplication and Data Routing Using a Partitioned Optical Passive Stars Network
IEEE Transactions on Parallel and Distributed Systems
The Partitioned Optical Passive Stars Network: Simulations and Fundamental Operations
IEEE Transactions on Parallel and Distributed Systems
The Partitioned Optical Passive Stars (POPS) topology
IPPS '95 Proceedings of the 9th International Symposium on Parallel Processing
Improved embeddings in POPS networks through stack-graph models
MPPOI '96 Proceedings of the 3rd Conference on Massively Parallel Processing Using Optical Interconnections
Fast merging and sorting on a partitioned optical passive stars network
Computer Science in Perspective
Hi-index | 0.00 |
In a Partitioned Optical Passive Stars (POPS) network, n = dg processors are divided into g groups of d processors each and such a POPS network is denoted by POPS(d, g). There is an optical passive star (OPS) coupler between every pair of groups. Each OPS coupler can receive an optical signal from any one of its source nodes and broadcast the signal to all the destination nodes. The time needed to perform this receive and broadcast is referred to as a time slot and the complexity of an algorithm using the POPS network is measured in terms of number of slots it uses. Since a POPS(d, g) requires g2 couplers, it is unlikely that in a practical system the number of couplers will be more than the number of processors. In other words, in most practical systems, the group size d will be greater than the number of groups g, i.e., d 驴n g. Hence, it is important to design fast algorithms for basic operations on such POPS networks with large group sizes. We present several fast algorithms for basic arithmetic operations on POPS(d, g)s such that d 驴n g. Our algorithms require significantly less number of slots for these operations compared to the best known algorithms for these problems designed by Sahni [8].