Computing Circuit Functional Safety Via Cross-Unit Result Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Chips face functional abnormalities in computing circuits due to factors like electromagnetic interference, which can compromise their reliability if not detected timely.
Innovation Solution
A functional safety protection method and computing circuit design that involves generating test data, performing computations across compute units, determining matching degrees of results, and using a multiplexer to facilitate functional safety detection during idle or operational states.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If functional safety detection is performed on computing circuits, then reliability is improved, but device complexity increases
Solution Approach 1:
The computing array is divided into multiple compute units that independently process test data. Each compute unit generates computation results that can be individually compared, allowing functional safety detection without requiring a complete system overhaul. This segmentation enables reliability monitoring while maintaining manageable complexity through modular architecture.
Solution Approach 2:
Test data is copied and distributed to multiple compute units for parallel processing. The same test data is processed independently by different compute units, and the results are compared to detect functional abnormalities. This copying approach enables safety detection without requiring additional hardware for each compute unit, as the same data is reused across multiple processing paths.
2Measurement precision
If multiple compute units perform computations to detect functional safety, then measurement precision is improved, but use of energy increases
Solution Approach 1:
The functional safety detection is performed periodically rather than continuously. Test data is generated and compute units are invoked to process the data at specific intervals, allowing the system to maintain measurement precision through regular sampling while reducing overall energy consumption compared to continuous monitoring. The system can adjust the frequency of detection based on operational requirements.
Solution Approach 2:
Instead of requiring all compute units to perform full computations continuously, the system invokes compute units to process test data only when functional safety detection is needed. This partial action approach maintains measurement precision by testing specific functions while minimizing energy consumption by keeping compute units in lower-power states during normal operation.
3Measurement precision
If compute units are invoked to perform computations for safety detection, then measurement precision is improved, but loss of time increases
Solution Approach 1:
Test data is generated and compute units are prepared in advance before functional safety detection is actually needed. The system can pre-compile test cases and pre-position compute units in a ready state, allowing rapid execution of safety detection when required. This preliminary preparation reduces the time overhead during actual detection operations while maintaining measurement precision.
Solution Approach 2:
The computing array continues to perform its primary computational functions while functional safety detection is conducted in parallel using the same compute units. By overlapping the safety detection process with normal operation rather than sequencing them exclusively, the system maintains measurement precision through continuous testing while minimizing time loss by utilizing otherwise idle computational cycles.
Data Source
AI summary
Disclosed are a functional safety protection method for a computing circuit and a functional safety computing circuit. The computing circuit includes a computing array with a plurality of compute units. The method includes: generating test data; invoking the compute units to respectively perform computations based on the test data, to obtain first computation results respectively corresponding to the compute units; determining a matching degree between the first computation results respectively corresponding to the compute units; and determining a functional safety detection result of the computing array based on the matching degree. According to embodiments of this disclosure, functional safety abnormalities that occur in the computing circuit can be detected, thereby facilitating protection for the computing circuit.