Plausibility-Driven Fault Detection in String Termination Logic
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional fault detection techniques in string search engines are inefficient and expensive, particularly for SIMD architectures, as they fail to effectively detect errors in substring search operations and do not provide adequate plausibility-driven cross-checking for internal signals and intermediate results.
Innovation Solution
A processor unit with vector registers and an M×M matrix of comparators performs character-wise comparisons between reference and target strings, generating zero detect vectors to identify terminating elements and resulting bit vectors indicating full or partial matches, enabling fault detection by comparing these vectors with operands.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional fault detection techniques are used in string search engines, then the system structure is simple, but fault detection efficiency is low and costs are high
Solution Approach 1:
The patent merges fault detection functionality with the existing string search engine by integrating zero detect vectors and plausibility checking into the comparator matrix architecture. The fault detection logic is combined with the substring search operations, allowing simultaneous execution of search and fault detection without requiring separate dedicated fault detection hardware, thus improving efficiency while maintaining reasonable system complexity
Solution Approach 2:
The patent introduces zero detect vectors as intermediary elements that serve dual purposes: they are used in the normal string matching process to identify terminating elements, and simultaneously serve as plausibility checks for fault detection. This intermediary mechanism enables efficient fault detection by reusing existing computational resources rather than adding separate detection infrastructure
2Reliability
If plausibility-driven cross-checking is implemented for internal signals, then reliability is improved, but device complexity increases
Solution Approach 1:
The system performs self-verification by using its own internal zero detect vectors to check for faults. The plausibility checking mechanism leverages the existing zero detect logic that is already part of the string search operation, allowing the system to self-diagnose faults without requiring external verification infrastructure. This self-service approach improves reliability while minimizing additional complexity
Solution Approach 2:
The zero detect vectors and comparator matrix serve multiple functions simultaneously: they perform the primary string matching operation, generate termination indicators, and provide plausibility checks for fault detection. This multi-functionality allows the same hardware components to support both normal operation and reliability verification, improving fault detection accuracy without proportionally increasing device complexity
3Reliability
If conventional fault detection methods are used, then implementation is simple, but they fail to detect errors in substring search operations
Solution Approach 1:
The patent performs preliminary plausibility checking by generating and validating zero detect vectors before they are used in the string matching process. The fault detection logic checks the consistency of these vectors in advance, identifying potential errors early in the computation pipeline. This preliminary action ensures that errors in substring search operations are detected while maintaining a relatively simple implementation by using basic vector comparison operations
Data Source
AI summary
A method for detecting faults in substring search operations using a processor unit including vector registers of M vector elements each. A non-limiting example of the method includes providing an M×M matrix of comparators for characterwise comparison of the elements of a reference string and a target string. A first zero detect vector having value indicative of terminating element of the target string and a second zero detect vector having a value indicative of terminating element of the reference string are generated. A resulting bit vector is generated using comparison performed by the M×M matrix. The resulting bit vector indicates characters of the target string that fully match the reference string and indicate characters of the target string that partially match the reference string. Fault detection in the substring search operations is performed by comparing the generated zero detect vectors with operands.