Arithmetic Processing Device Pipeline Stage Priority Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In pipeline parallel systems, stages with longer processing times can lead to decreased throughput due to data accumulation and resource conflicts, as resources are not efficiently allocated based on actual processing times.
Innovation Solution
An arithmetic processing device measures the processing time of each pipeline stage and sets priorities in descending order, allowing for efficient resource allocation and preventing stages with longer processing times from being interrupted, thereby equalizing processing times and improving throughput.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If pipeline parallel processing is implemented to improve throughput, then processing speed is improved, but resource conflicts and data accumulation occur when stages have unequal processing times
Solution Approach 1:
The system dynamically adjusts the priority of pipeline stages based on their actual processing times. The priority setting unit changes the execution priority of each stage according to measured performance data, making the pipeline adaptive to varying processing requirements and preventing bottlenecks caused by static priority assignments.
Solution Approach 2:
The measurement unit continuously monitors the processing time of each pipeline stage and feeds this information back to the priority setting unit. This feedback loop enables the system to automatically adjust priorities based on actual performance, ensuring that stages with longer processing times receive higher priority to maintain balanced throughput.
2Device complexity
If resources are allocated equally to all pipeline stages, then resource allocation is simple, but stages with longer processing times cause data accumulation and reduce overall efficiency
Solution Approach 1:
Instead of uniform resource allocation, the system applies differentiated priority levels to different pipeline stages based on their individual processing characteristics. Each stage receives resources proportional to its needs, with stages having longer processing times allocated higher priority, thereby optimizing overall pipeline efficiency without excessive complexity.
3Ease of operation
If pipeline stages are executed in parallel without priority differentiation, then ease of operation is maintained, but resource conflicts occur reducing processing efficiency
Solution Approach 1:
The pipeline system automatically determines and adjusts its own priorities based on measured processing times. The measurement unit and priority setting unit work autonomously to optimize resource allocation without requiring external intervention or complex manual configuration, maintaining ease of operation while improving processing efficiency through adaptive priority management.
Data Source
AI summary
An arithmetic processing device includes: a memory; and a processor coupled to the memory and configured to: execute a plurality of data processes each of which is divided into a plurality of pipeline stages in parallel at different timings; measure a processing time of each of the plurality of pipeline stages; and set a priority of the plurality of pipeline stages in a descending order of the measured processing time.