Fat-trees and low dimensional cubes have raised a great interest in the scientific community in the last few years and are emerging standards in the design of interconnection networks for massively parallel computers. In this paper we compare the communication performance of these two classes of interconnection networks using a detailed simulation model which describes both the interconnection network and the memory hierarchy of the processing nodes. The comparison is made with a real parallel algorithm, the transpose FFT algorithm, taking into account physical constraints, as pin and bandwidth limitations, and the router complexity
A central issue in the design of a general-purpose parallel computer is the choice of an interconne...
The computation of a one-dimensional FFT on a c-dimensional torus multicomputer is analyzed. Differe...
Communication with uniformly distributed destinations, with locality and with hot-spots are simulat...
This paper provides empirical comparison of the communication capabilities of two area-universal net...
2D-mesh and torus networks have often been proposed as the interconnection pattern for parallel comp...
2D-mesh and torus networks have often been proposed as the interconnection pattern for parallel comp...
This paper compares message routing capabilities of important networks proposed for general-purpose ...
Many theoretical models of parallel computation are based on overly simplistic assumptions on th...
Network line, cards are experiencing ever increasing line rates, random data bursts, and limited spa...
Network line cards are experiencing ever increas-ing line rates, random data bursts, and limited spa...
The topology of interconnection networks the stage a key role in the performance of all general purp...
A recent renewed interest in hypercube interconnection network has been concentrated to the more sca...
This paper shows that a novel network called the fat-stack is universally efficient when adequate ca...
Many applications have stimulated the recent surge of interest single-chip parallel processing. In s...
Wormhole routing and deflection routing are two of the most promising techniques used for routing me...
A central issue in the design of a general-purpose parallel computer is the choice of an interconne...
The computation of a one-dimensional FFT on a c-dimensional torus multicomputer is analyzed. Differe...
Communication with uniformly distributed destinations, with locality and with hot-spots are simulat...
This paper provides empirical comparison of the communication capabilities of two area-universal net...
2D-mesh and torus networks have often been proposed as the interconnection pattern for parallel comp...
2D-mesh and torus networks have often been proposed as the interconnection pattern for parallel comp...
This paper compares message routing capabilities of important networks proposed for general-purpose ...
Many theoretical models of parallel computation are based on overly simplistic assumptions on th...
Network line, cards are experiencing ever increasing line rates, random data bursts, and limited spa...
Network line cards are experiencing ever increas-ing line rates, random data bursts, and limited spa...
The topology of interconnection networks the stage a key role in the performance of all general purp...
A recent renewed interest in hypercube interconnection network has been concentrated to the more sca...
This paper shows that a novel network called the fat-stack is universally efficient when adequate ca...
Many applications have stimulated the recent surge of interest single-chip parallel processing. In s...
Wormhole routing and deflection routing are two of the most promising techniques used for routing me...
A central issue in the design of a general-purpose parallel computer is the choice of an interconne...
The computation of a one-dimensional FFT on a c-dimensional torus multicomputer is analyzed. Differe...
Communication with uniformly distributed destinations, with locality and with hot-spots are simulat...