Relay Control Node for Parallel Calculation Route Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In large-scale parallel calculation systems, determining an inter-node communication route that minimizes interference between jobs is challenging due to the complexity of calculating interference costs across numerous potential relay nodes, especially in systems using dimension-order routing, which increases calculation time and makes it difficult to select relay nodes efficiently.
Innovation Solution
The implementation of a relay control node that generates a physical network graph, hierarchizes the minimum cost flow problem, and specifies a communication route with the lowest passing cost by using an interconnected component graph, reducing the number of edges and vertices to minimize interference and calculation time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If dimension-order routing is used to determine communication routes, then route determination is automated by hardware, but calculation time increases and it becomes difficult to select relay nodes efficiently to minimize interference between jobs
Solution Approach 1:
The patent segments the mesh network into multiple sub-meshes, each assigned to different jobs. This segmentation allows independent route determination within each sub-mesh, reducing the overall calculation complexity and time while maintaining automated hardware-based routing within each segment.
Solution Approach 2:
The patent introduces a hierarchical dimension to the routing problem by creating sub-meshes within the overall mesh network. This dimensional transformation allows the system to solve the routing problem at multiple levels (sub-mesh level and overall network level), reducing the computational burden at each level.
2Adaptability or versatility
If multiple sub-meshes are used by one job to increase node availability, then node utilization improves, but communication interference between jobs occurs
Solution Approach 1:
The patent applies local quality by assigning different characteristics to different sub-meshes. Each sub-mesh can be independently configured and assigned to specific jobs, allowing the system to optimize node availability for each job while preventing interference through spatial separation. Jobs can access multiple nodes within their assigned sub-mesh without interfering with other jobs.
3Object-generated harmful factors
If relay nodes are set to avoid communication interference, then interference between jobs is reduced, but the complexity of calculating interference costs across numerous potential relay nodes increases
Solution Approach 1:
The patent performs preliminary actions by pre-calculating and storing interference cost information for different relay node selections. This pre-computation allows the system to quickly determine optimal relay nodes during runtime without performing complex real-time calculations, thus reducing communication interference while managing calculation complexity.
Solution Approach 2:
The patent implements dynamic relay node selection based on current network conditions and job requirements. The system can adaptively choose relay nodes by considering real-time factors such as current traffic patterns and job priorities, allowing flexible interference avoidance without requiring exhaustive calculation of all possible relay node combinations.
Data Source
AI summary
An information processing apparatus is configured to receive a request for communication between a first node and a second node included in a parallel calculation system, acquire job execution information relating a job to be executed by the parallel calculation system, generate connected graph information based on first information on the first node, second information on the second node, the job execution information, and topology information indicating a topology of the plurality of nodes, generate, based on the connected graph information, route information indicating a plurality of routes used when the communication between the first node and the second node is executed, specify, based on the route information, a route having the lowest passing cost among the plurality of routes; and specify a node included in the specified route as a relay node based on positions of the plurality of nodes in the specified route.