Method for simulataneous overlapped communications between neighboring processors in a multiple
This patent describes a method of operating a computing system having a MIMD architecture where a program is concurrently run on a plurality of nodes for reducing total processing time for the program; each node having a processor, a memory, and a communication channel connected to the each node and independently connected to a predetermined number of other nodes; it comprises: assigning, via the computing system, groups of adjacent nodes, each the group having communication channels from a central node to four adjacent nodes in up, down, left, and right directions, respectively, each of the nodes being capable of being a central node in a different group; exchanging, via the computing system, messages for each of the nodes; and simultaneously communicating, via the computing system, the messages between adjacent nodes.
- Assignee:
- Dept. of Energy, Washington, DC (United States)
- Patent Number(s):
- US 5 072371; A
- Application Number:
- PPN: US 7-317389
- OSTI ID:
- 6885238
- Resource Relation:
- Patent File Date: 1 Mar 1989
- Country of Publication:
- United States
- Language:
- English
Similar Records
Simulation study of a parallel processor with unbalanced loads. Master's thesis
Methods for operating parallel computing systems employing sequenced communications