Multi-Core I/O Processing with Dynamic Polling and Interrupt Switching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current I/O processing methods either lead to high CPU load and decreased efficiency due to repetitive polling or slow I/O processing due to CPU interrupts, as they do not effectively balance between polling and interrupt-based detection of reception events.
Innovation Solution
An information processing device with multiple processor cores that execute processes based on priority levels, using polling when possible and CPU interrupts when necessary, to optimize I/O processing speed and efficiency by dynamically switching between these methods based on load conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If polling processing is used to detect reception events, then I/O processing speed is improved, but CPU load increases and efficiency of other processing decreases
Solution Approach 1:
The system dynamically switches between polling and interrupt-based detection methods based on the execution state of polling processes. When polling is active, interrupt detection is suspended to avoid context switches; when polling is inactive, interrupt detection is enabled to reduce CPU load. This dynamic adaptation resolves the contradiction between I/O speed and CPU efficiency.
2Productivity
If CPU interrupt is used to detect reception events, then CPU load is reduced and efficiency of other processing improves, but I/O processing speed decreases due to context switch overhead
Solution Approach 1:
The system dynamically switches between polling and interrupt-based detection methods based on the execution state of polling processes. When polling is active, interrupt detection is suspended to avoid context switches; when polling is inactive, interrupt detection is enabled to reduce CPU load. This dynamic adaptation resolves the contradiction between I/O speed and CPU efficiency.
3Reliability
If polling is executed repeatedly by the CPU, then reception events are detected without context switches, but CPU load increases due to repetitive processing
Solution Approach 1:
The system dynamically switches between polling and interrupt-based detection methods based on the execution state of polling processes. When polling is active, interrupt detection is suspended to avoid context switches; when polling is inactive, interrupt detection is enabled to reduce CPU load. This dynamic adaptation resolves the contradiction between I/O speed and CPU efficiency.
Solution Approach 2:
The system uses a polling process as an intermediary mechanism to detect reception events. Instead of having the CPU directly execute repetitive polling instructions, a scheduled polling process performs the detection work, allowing the CPU to switch to other tasks when polling is not needed. This intermediary approach reduces direct CPU load while maintaining reliable event detection.
Data Source
AI summary
An information processing device includes a memory, and a plurality of processor cores that access the memory. The plurality of processor cores respectively executes processes to be executed by the plurality of processor cores in accordance with execution priority levels of the processes. When a polling process for repeatedly determining whether reception data for input/output processing is received is underway in one of the plurality of processor cores, the plurality of processor cores respectively executes the input/output processing in response to a determination, made by the polling process, that the reception data have been received, and when the polling process is not underway in any of the plurality of processing cores, the plurality of processor cores respectively executes the input/output processing in response to a processor interrupt issued upon reception of the reception data.