Allocating application to group of consecutive processors in fault-tolerant deadlock-free routing path defined by routers obeying same rules for path selection
View Patent ↗In a multiple processor computing apparatus, directional routing restrictions and a logical channel construct permit fault tolerant, deadlock-free routing. Processor allocation can be performed by creating a linear ordering of the processors based on routing rules used for routing communications between the processors. The linear ordering can assume a loop configuration, and bin-packing is applied to this loop configuration. The interconnection of the processors can be conceptualized as a generally rectangular 3-dimensional grid, and the MC allocation algorithm is applied with respect to the 3-dimensional grid.
1. A computing apparatus, comprising:
a plurality of processors;
a communication network structure including a plurality of routers that are respectively connected to respectively associated ones of said plurality of processors, each of said routers having associated therewith a plurality of communication paths that connect said router to others of said routers, said processors interconnected for communication with one another by said routers and said communication paths, each of said routers configured to implement a same set of routing rules to determine which of the associated communication paths will be used to forward a communication that is received by said router while enroute between a communicating pair of said processors; and
a processor allocator coupled to said communication network structure and configured to select a group of said processors to cooperatively execute an application;
wherein said processor allocator selects said group of processors based on a mapping between said plurality of processors and a plurality of respectively associated identifiers that identify the associated processors and define a sequence;
wherein said processor allocator selects said group of processors by selecting a group of said identifiers whose respectively associated processors have processing capacity available to support execution of said application; and
wherein said mapping is based on said routing rules.
2. The apparatus of claim 1 , wherein said mapping defines a corresponding relationship between said plurality of routers and respectively associated ones of said identifiers.
3. The apparatus of claim 2 , wherein adjacent pairs of said identifiers in said sequence respectively correspond to pairs of said routers that are only one communication hop apart in said communication network structure.
4. The apparatus of claim 3 , wherein first and last ones of said identifiers in said sequence correspond to respective ones of said routers that are only one communication hop apart in said communication network structure.
5. The apparatus of claim 2 , wherein first and last ones of said identifiers in said sequence correspond to respective ones of said routers that are only one communication hop apart in said communication network structure.