Shared Cache Split and Lock Modes for Safety-Critical Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data processing systems in safety-critical applications, such as automotive systems, face challenges in balancing performance and functional safety, particularly in managing shared caches across different operational modes to ensure error detection without sacrificing cache performance or incurring additional hardware costs.
Innovation Solution
A data processing apparatus with a cluster of processing elements that operates in both split and lock modes, where a shared cache with full capacity is accessible in both modes, allowing primary processing elements to perform workloads while redundant elements check for errors without needing additional redundant RAM cells or chopped cache configurations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a shared cache is divided into portions for primary and redundant processing elements in lock mode, then functional safety is improved through error detection, but cache performance is worsened due to reduced accessible cache capacity
Solution Approach 1:
The cache allocation is made dynamic rather than static. In split mode, the cache is divided into portions for each processing element. In lock mode, the cache dynamically switches to full accessibility for primary processing elements, allowing the system to adapt cache resources based on operational mode requirements.
Solution Approach 2:
The cache accessibility parameter changes based on operational mode. The system transitions from a state where cache is partitioned (split mode) to a state where full cache capacity is accessible (lock mode), changing the cache configuration parameter to optimize for either parallel processing or functional safety verification.
2Reliability
If additional redundant RAM cells are added for lock mode operation, then functional safety is improved through error detection, but device complexity and hardware cost are worsened
Solution Approach 1:
The shared cache serves multiple functions across different operational modes. The same cache infrastructure is used for both split mode operation (with partitioned access) and lock mode operation (with full access for primary elements), eliminating the need for separate redundant cache structures and reducing overall hardware complexity.
Solution Approach 2:
The patent merges the cache resources into a single shared structure that serves both primary and redundant processing elements. Instead of maintaining separate cache memories for redundancy, the system combines cache resources into one shared pool that can be allocated differently based on operational mode, reducing total hardware requirements.
3Reliability
If the shared cache capacity is reduced to accommodate redundant elements, then functional safety is improved, but cache performance and accessibility are worsened
Solution Approach 1:
The cache capacity allocation is dynamic rather than fixed. The system can utilize the full cache capacity in lock mode when functional safety verification is performed, and partition the same capacity in split mode for parallel processing, maximizing resource utilization without requiring additional hardware.
Data Source
AI summary
A cluster of processing elements has a split mode in which the processing elements are configured to process independent processing workloads, and a lock mode in which the processing elements comprise at least one primary processing element and at least one redundant processing element, each redundant processing element configured to perform a redundant processing workload for checking an outcome of a primary processing workload performed by a corresponding primary processing element. A shared cache is provided, having a predetermined cache capacity accessible to each of the processing elements when in the split mode. In the lock mode, the predetermined cache capacity of the shared cache is fully accessible to the at least one primary processing element.