Implementation Method for Parallel Computing in Multi-Cluster Grid
XU Fu-yuanb
Abstract
XU Fu-yuanb
Abstract
For implementing DPC(Data Parallel Computing) in multi-cluster grid that composed of many computer clusters,a series of formal definitions,such as multi-cluster grid,logical computer cluster and three kinds of agents were given.By using a cooperation mechanism among the management Agent,the cooperation computing Agent,the communication agent and the cooperation computing team,the DPC process were studied.Based on the state of computer,such as main state,redundancy state and failure state,dynamic resource partition and grid computing process were described.The experiments show that this method can effectively adapt to the heterogeneous and dynamic characteristic in multi-cluster grid.The speed up of parallel computing by this method is approximately linear.It can be fit for the parallel computing in gird.
A significance statement is not available in the OpenAlex record.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
For implementing DPC(Data Parallel Computing) in multi-cluster grid that composed of many computer clusters,a series of formal definitions,such as multi-cluster grid,logical computer cluster and three kinds of agents were given.By using a cooperation mechanism among the management Agent,the cooperation computing Agent,the communication agent and the cooperation computing team,the DPC process were studied.Based on the state of computer,such as main state,redundancy state and failure state,dynamic resource partition and grid computing process were described.The experiments show that this method can effectively adapt to the heterogeneous and dynamic characteristic in multi-cluster grid.The speed up of parallel computing by this method is approximately linear.It can be fit for the parallel computing in gird.
Key concepts: Grid computing, Computer science, Computer cluster, Distributed computing, Grid, Partition (number theory), Cluster (spacecraft), Redundancy (engineering)