Communication effect basic linear algebra computations on hypercube architectures
Journal of Parallel and Distributed Computing
Nearest-neighbor mapping of finite element graphs onto processor meshes
IEEE Transactions on Computers
Introduction to VLSI Systems
Embedding Rectangular Grids into Square Grids
AWOC '88 Proceedings of the 3rd Aegean Workshop on Computing: VLSI Algorithms and Architectures
On the Assignment Problem of Arbitrary Process Systems to Heterogeneous Distributed Computer Systems
IEEE Transactions on Computers
An algebraic theory for modeling direct interconnection networks
Proceedings of the 1992 ACM/IEEE conference on Supercomputing
Realizing Common Communication Patterns in Partitioned Optical Passive Stars (POPS) Networks
IEEE Transactions on Computers
Embedding Binary Trees into Crossed Cubes
IEEE Transactions on Computers
Optimal Processor Assignment for a Class of Pipelined Computations
IEEE Transactions on Parallel and Distributed Systems
Topology mapping for Blue Gene/L supercomputer
Proceedings of the 2006 ACM/IEEE conference on Supercomputing
| Hi-index | 14.99 |
A novel technique, the multiple ripple propagation technique, is presented for mapping and h*w grid into a w*h grid such that the dilation cost is 2, i.e. such that any two neighboring nodes in the first grid are mapped onto two nodes in the second grid that are separated by a distance of at most 2. The technique is then used as a basic tool for mapping any rectangular source grid into a square target grid with the dilation two property preserved. The ratio of the number of nodes in the source grid to the number of nodes in the target grid, called the expansion cost, is shown to be always less than 1.2. This is a significant improvement over the previously suggested techniques, where the expansion cost could be bounded by 1.2 only if the dilation cost was allowed to be as high as 18.