Work Request Processor Scheduling via On-Deck Unit
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing network processors face inefficiencies in scheduling and queuing work requests for multiple processors, leading to suboptimal utilization and waiting times due to shared work queues where only the first processor executes a task, causing others to wait for additional jobs.
Innovation Solution
A system utilizing an add work engine to forward work queue entries to input queues and an on-deck unit that schedules work by prioritizing and distributing work queue entries among multiple processors, ensuring high-frequency scheduling and maximizing processor capacity through a centralized table and list management system.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a shared work queue is used where only the first processor executes a task, then the work queue structure is simple, but processor utilization is low and waiting time increases
Solution Approach 1:
The patent segments the work queue into multiple independent queues, one for each processor. Each processor has its own dedicated work queue, allowing simultaneous processing without interference. This segmentation resolves the contradiction by maintaining simple queue structures while enabling all processors to be utilized effectively, eliminating the waiting time problem of shared queues.
Solution Approach 2:
The patent transitions from a single-dimensional shared work queue to a multi-dimensional structure where each processor has its own queue. This dimensional expansion allows parallel processing paths, improving processor utilization while keeping each individual queue simple in structure.
2Productivity
If work queue entries are distributed among multiple processors, then processor capacity is maximized, but scheduling complexity increases
Solution Approach 1:
The scheduling system is segmented into simple, independent work queue management for each processor. Instead of a complex centralized scheduler, each processor manages its own queue independently, reducing scheduling complexity while maximizing processor capacity through parallel processing.
Solution Approach 2:
Each processor independently manages its own work queue without requiring complex centralized scheduling. The processors self-manage their task assignments and execution, simplifying the overall scheduling system while enabling efficient parallel processing across multiple processors.
3Ease of operation
If multiple processors wait for additional jobs in a shared queue, then work distribution is simple, but waiting time increases
Solution Approach 1:
The work distribution system is segmented into separate queues for each processor, allowing simultaneous progress on different tasks. This eliminates the waiting time problem while maintaining simple work distribution mechanisms, as each processor independently receives and processes tasks from its dedicated queue.
Solution Approach 2:
The patent enables continuous useful action for all processors by providing dedicated work queues. Processors continuously process tasks from their respective queues without waiting for additional jobs, maximizing productivity while keeping work distribution simple and straightforward.
Data Source
AI summary
A network processor includes a schedule, sync and order (SSO) module for scheduling and assigning work to multiple processors. The SSO includes an on-deck unit (ODU) that provides a table having several entries, each entry storing a respective work queue entry, and a number of lists. Each of the lists may be associated with a respective processor configured to execute the work, and includes pointers to entries in the table. pointer is added to the list based on an indication of whether the associated processor accepts the WQE corresponding to the pointer.