Processor Thread Isolation for Core Error Containment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing computing systems face inefficiencies in handling hardware faults, particularly processor core errors, leading to unnecessary shutdowns due to uncorrupted shared system resources, which can disrupt unaffected workloads.
Innovation Solution
Implementing core isolation by marking a thread as isolated upon error detection, using a control circuit to notify a partner thread and prevent further execution, thereby avoiding system shutdowns and maintaining unaffected workloads.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If system shutdown is implemented in response to hardware errors, then system reliability is improved, but productivity deteriorates due to unnecessary shutdowns disrupting unaffected workloads
Solution Approach 1:
The patent segments the system response by isolating the faulty thread from the rest of the system. Instead of shutting down the entire system, only the isolated thread is stopped and marked, allowing other threads to continue executing unaffected workloads. This segmentation resolves the contradiction by maintaining system reliability through error isolation while preserving productivity through continued operation of healthy threads.
Solution Approach 2:
The patent applies local quality by treating the faulty thread differently from healthy threads. The isolated thread is marked with a special status and prevented from executing, while other threads continue normal execution. This localized response ensures system reliability for the faulty component without impacting overall system productivity.
2Productivity
If core isolation is implemented to prevent error spread, then productivity is improved by avoiding shutdowns, but device complexity increases due to additional isolation mechanisms
Solution Approach 1:
The thread isolation mechanism is self-service in nature, where the thread itself is marked as isolated through its own error conditions. The control circuit automatically identifies and isolates faulty threads without requiring external intervention or complex manual configuration. This self-service approach improves productivity through automatic error containment while minimizing the increase in device complexity.
Solution Approach 2:
The control circuit acts as an intermediary between the error detection mechanism and the thread execution. It receives error signals, determines thread isolation, and manages the isolation process. This intermediary approach simplifies the overall system by centralizing isolation logic rather than requiring complex distributed decision-making across all threads.
3Reliability
If thread isolation is implemented to contain errors, then reliability is improved by preventing error propagation, but loss of time increases due to thread isolation overhead
Solution Approach 1:
The patent implements skipping by rapidly marking the faulty thread as isolated and preventing further execution without engaging in lengthy error analysis or system-wide shutdown procedures. The control circuit quickly identifies the problematic thread and isolates it, minimizing the time lost while ensuring reliability through proper error containment.
Solution Approach 2:
The patent applies partial action by isolating only the necessary thread components required for error containment rather than shutting down the entire system. This partial isolation approach achieves reliability through targeted error containment while minimizing time loss by avoiding unnecessary system-wide interruptions.
Data Source
AI summary
The disclosed device includes a controller that can receive an error interrupt for a thread, and isolate the thread. The controller can also notify a partner thread of the isolated thread. The partner thread can report the isolation to an operating system or other controller. Isolating the thread can avoid halting other unaffected threads. Various other methods, systems, and computer-readable media are also disclosed.