Approximate Calculation Integrity Checking for Processor Fault Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for detecting intermittent faults in processing circuitry, such as mercurial cores, are inefficient and costly, particularly when redundant execution or dual core lock step approaches are used, and there is a need for a more efficient method to check the integrity of calculations.
Innovation Solution
Incorporating approximation circuitry to calculate an approximate result of a calculation concurrently with the main processing circuitry, and comparing this result with a deviation threshold to detect errors in the processing circuitry, thereby reducing the need for redundant calculations or duplicate cores.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If redundant execution or dual core lock step approaches are used to detect intermittent faults, then detection reliability is improved, but device complexity and power consumption increase
Solution Approach 1:
The patent creates a simplified copy of the calculation circuitry called approximation circuitry that performs approximate calculations instead of exact calculations. This copy is sufficient for integrity checking purposes but consumes fewer resources than full redundant execution or dual core approaches, thereby resolving the contradiction between detection reliability and device complexity
Solution Approach 2:
The patent changes the precision parameter of the calculation circuitry by introducing approximation circuitry that calculates results with reduced precision. By comparing the approximate result with the exact result and checking if their difference exceeds a threshold, the system achieves fault detection without requiring full redundant circuitry, thus reducing device complexity while maintaining reliability
2Reliability
If redundant execution or dual core lock step approaches are used to detect intermittent faults, then detection reliability is improved, but power consumption increases
Solution Approach 1:
The approximation circuitry serves as a simplified copy that performs only approximate calculations, consuming significantly less power than full redundant execution or dual core lock step approaches while still enabling effective fault detection through result comparison
Solution Approach 2:
The patent applies partial action by implementing only the necessary portion of redundant calculation (approximate calculation) rather than full redundancy. The approximation circuitry performs sufficient calculation to detect faults through threshold comparison without the excessive power consumption of complete redundant systems
3Use of energy by moving object
If approximation circuitry is used to detect errors, then power consumption and area are reduced, but measurement precision decreases
Solution Approach 1:
The patent deliberately changes the precision parameter by using approximation circuitry with reduced precision. However, this is acceptable because the system compares the approximate result with the exact result and uses a threshold to detect significant deviations, thereby maintaining effective error detection while reducing power consumption and area
Data Source
AI summary
An apparatus has processing circuitry to execute instructions. The processing circuitry has calculation circuitry which is responsive to one or more instructions requiring a calculation to be performed to compute the result of the calculation and approximation circuitry which is responsive to said one or more instructions to calculate an approximate result of the calculation independently of the calculation circuitry. The processing circuitry also has integrity checking circuitry to perform an integrity check by comparing the result of the calculation performed by the calculation circuitry and the approximate result of the calculation performed by the approximation circuity. The integrity checking circuitry detects an error in the processing circuitry if it is determined that a difference between the result of the calculation and the approximate result of the calculation is greater than a deviation threshold.