Graphics Task Repetition Across Units for Fault Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing graphics processing units (GPUs) in safety-critical systems face challenges in meeting stringent safety standards like ASIL D of ISO 26262 due to high power and area consumption in dual lockstep processors, which are not suitable for advanced driver-assistance systems and autonomous vehicles that require significant graphics and vector processing capabilities.
Innovation Solution
A graphics processing system with multiple independent processing units that process tasks twice, forming signatures for comparison to detect faults, allowing efficient fault detection without doubling chip area or power consumption, and distinguishing between safety-critical and non-safety-critical tasks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If dual lockstep processors are used to meet safety standards, then fault detection capability is improved, but chip area and power consumption double
Solution Approach 1:
The system segments fault detection into two independent processing units that operate separately but can be compared. Each unit processes tasks independently, and fault detection is achieved through comparison of their outputs rather than requiring both units to operate simultaneously in lockstep mode, thereby reducing the area required for redundant processing.
Solution Approach 2:
Instead of requiring full lockstep redundancy for all processing operations, the system applies partial redundancy by processing certain safety-critical tasks twice through different processing units and only comparing results when necessary. This selective approach reduces overall chip area while maintaining safety capabilities for critical functions.
2Reliability
If dual lockstep processors are used to meet safety standards, then fault detection capability is improved, but power consumption doubles
Solution Approach 1:
The processing system is segmented into multiple independent processing units that can operate independently. Safety-critical tasks are processed twice through different units, but the system can optimize power consumption by controlling which units are active based on task requirements, rather than maintaining continuous lockstep operation of all units.
Solution Approach 2:
The system employs periodic comparison of processing results rather than continuous lockstep operation. Tasks are processed, results are compared, and only then is fault detection activated. This periodic approach maintains safety capabilities while significantly reducing average power consumption compared to continuous dual lockstep processing.
3Reliability
If all tasks are processed twice for safety, then fault detection is improved, but processing efficiency deteriorates
Solution Approach 1:
The system applies different processing strategies to different task types. Safety-critical tasks are processed twice through different processing units for fault detection, while non-safety-critical tasks are processed once. This local differentiation of processing quality ensures fault detection for critical functions while maintaining overall processing efficiency.
Solution Approach 2:
Instead of processing all tasks twice, the system applies partial redundancy only to safety-critical tasks. Non-critical tasks are processed once, and the system uses task classification to determine which tasks require duplicate processing. This selective approach maintains fault detection capability for critical functions while optimizing overall processing efficiency.
Data Source
AI summary
A graphics processing system includes a plurality of processing units, wherein the graphics processing system is configured to process a task first and second times at the plurality of processing units. Data identifying which processing unit of the plurality of processing units the task has been allocated to is consulted on allocating the task to a processing unit for processing for a second time, and, in response, the task is allocated for processing for the second time to any processing unit of the plurality of processing units other than the processing unit to which the task was allocated for processing for a first time.