Distributed Tag Matching for Wildcard MPI Receives
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Distributed tag matching in Message Passing Interface (MPI) systems is inefficient when wild-carded data sources are involved, as existing methods require all matching to be centralized, reducing the efficiency of distributed approaches.
Innovation Solution
A distributed matching system that processes wildcarded receives by distributing the matching to multiple engines, using a master matching table to track wild-carded receive tags and ensuring each tag is matched only once, with parallel processing across multiple engines.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If all tag-matching is centralized in a single matching engine, then tag matching reliability is improved, but system efficiency and scalability deteriorate
Solution Approach 1:
The patent divides the centralized matching engine into multiple distributed matching engines that share the tag matching workload. Each engine maintains a portion of the unmatched message list and processes messages independently, enabling parallel processing while maintaining the reliability of complete matching through coordination mechanisms.
2Adaptability or versatility
If wild-carded receive tags are distributed to multiple matching engines, then system scalability is improved, but matching correctness deteriorates due to potential duplicate matching
Solution Approach 1:
The patent implements a token-based preliminary action mechanism where a token manager issues unique tokens to matching engines before they begin processing wild-carded receive tags. This preliminary token allocation ensures that each engine knows which tags it is responsible for, preventing duplicate matching while enabling distributed processing.
Solution Approach 2:
The patent employs feedback mechanisms where matching engines report their matching progress and results back to the token manager and other engines. This feedback loop enables coordination to ensure each wild-carded receive tag is matched exactly once across the distributed system, maintaining correctness while scaling to multiple engines.
3Device complexity
If wild-carded receive tags are processed sequentially in a single engine, then matching simplicity is maintained, but processing time increases
Solution Approach 1:
The patent segments the processing of wild-carded receive tags across multiple matching engines, with each engine handling a subset of tags simultaneously. This segmentation enables parallel processing that reduces total processing time while maintaining algorithmic simplicity at each individual engine.
Solution Approach 2:
The patent introduces a temporal dimension to the processing model by enabling asynchronous parallel execution across multiple engines. Instead of sequential processing in one dimension, the system processes tags across multiple engines simultaneously, adding a time dimension that reduces overall processing duration.
Data Source
AI summary
Embodiments of the present disclosure are directed to distributed matching of messages to an intended recipients when the message is processed as a wildcard receive by a recipient node. Generally speaking, embodiments described herein address distributed tag-matching in the presence of matching tags with wild-carded receives and that considers parameters not associated with the matching-tag which are not known to both the sender and receiver a priori.