Lockstep Processor Error Detection with Adaptive BIST Recovery

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing lockstep computing systems face challenges in efficiently and quickly identifying and recovering from both soft and hard errors in multiple processing cores, which can lead to system instability and potential hazards.

Innovation Solution

A lockstep processing system employs a built-in self-test (BIST) procedure to differentiate between soft and hard errors by analyzing output signals from multiple processing cores, adaptively probing likely error sources, and implementing recovery strategies based on error type, including resetting or disabling faulty cores to ensure safe operation.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If a redundant processor is added to detect errors, then reliability is improved, but device complexity increases

Engineering Contradiction:
Improveerror detection capabilityVSAvoidprocessor configuration
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent combines the redundant processor with the main processor into a single integrated unit, where the redundant processor is configured to execute the same instructions as the main processor. This merging approach allows error detection functionality to be incorporated without requiring a completely separate detection system, thereby improving reliability while limiting the increase in device complexity through shared hardware resources and unified instruction execution.

Inventive Principle:
Principle #5Merging (Combining)

2Measurement precision

If the redundant processor executes all instructions, then measurement precision is improved, but use of energy increases

Engineering Contradiction:
Improveinstruction execution accuracyVSAvoidprocessor energy consumption
Core Design Contradiction:
Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The patent implements partial execution by configuring the redundant processor to execute only a subset of instructions rather than all instructions. The system selectively executes instructions based on detected errors or specific conditions, thereby maintaining measurement precision for critical operations while reducing overall energy consumption by avoiding unnecessary execution of all instructions whenever errors are detected.

Inventive Principle:
Principle #16Partial or excessive action

Solution Approach 2:

The redundant processor incorporates feedback mechanisms that monitor the execution status and error conditions of instructions. Based on this feedback, the system dynamically adjusts the execution strategy by selecting which instructions to execute and which to skip, optimizing the balance between measurement precision and energy consumption according to the actual operational state.

Inventive Principle:
Principle #23Feedback

Data Source

PatentEP3776210B1Device, system and process for redundant processor error detection
Publication Date: 2026.05.20 ARM LTD
  • EP3776210B1 patent drawingFigure 1
  • EP3776210B1 patent drawingFigure 2A
  • EP3776210B1 patent drawingFigure 2B

AI summary

Briefly, example methods, apparatuses, and/or articles of manufacture are disclosed that may be implemented, in whole or in part, to determine indicators of potential errors in a multi- processing core lockstep computing device comprising a plurality of processing cores, based, at least in part, on observations of output signals generated by at least two processing cores of the plurality of processing cores. A built-in self-test (BIST) procedure may then be based, at least in part, on the determining indicators.