Interprocess resource-based dynamic scheduling system and method
A method and system for scheduling tasks in a processing system. In one embodiment, the method comprises processing tasks from a primary work queue, wherein the tasks consume resources that are operable to be released. Whenever the volume of resources that have been consumed exceeds a threshold, the processor executes tasks from a secondary work queue for a period of time. The secondary work queue is comprised of tasks from the primary work queue that can release the resources; the secondary work queue can be sorted according to the volume of resources that can be released.
1. A method of scheduling tasks in a computer system, comprising:
populating a primary work queue with a plurality of tasks;
processing said plurality of tasks from said primary work queue, wherein said plurality of tasks consume interprocess communication resources operable to be released;
monitoring said interprocess communication resources consumed by each task in said primary work queue;
determining whether a volume of consumed interprocess communication resources exceeds a threshold based on said monitoring;
processing tasks from a secondary work queue for a given time period, wherein said secondary work queue comprises tasks from said primary work queue that can release the greatest volume of interprocess communication resources, and wherein said tasks in said secondary work queue are sorted according to a respective priority value, wherein said threshold and said given time period are configurable at system initialization.
2. The method as recited in claim 1 , further comprising monitoring the interprocess communication resources that can be released by each task in said primary work queue.
3. The method as recited in claim 1 , wherein said tasks in said secondary work queue are further sorted according to a respective volume of interprocess communication resources that can be released.
4. The method as recited in claim 1 , wherein said secondary work queue is populated prior to execution from said secondary work queue.
5. The method as recited in claim 1 , wherein said primary work queue is executed according to preemptive priority scheduling.
6. The method as recited in claim 1 , wherein a given task that has received processing time in said primary work queue is not used to populate said secondary work queue.
7. The method as recited in claim 1 , wherein said interprocess communication resources are chosen from at least one of shared memory, semaphores, message queues, message pipes, sockets, and signals.
8. A method for scheduling tasks in a processor, the method comprising:
populating a primary work queue with a plurality of tasks;
processing said plurality of tasks from said primary work queue for a first time period;
during execution of tasks from said primary work queue, maintaining a list of interprocess communication resources pending for each task in said primary work queue;
determining, after said first time period, whether said interprocess communication resources pending has exceeded a threshold based on said maintained list of interprocess communication resources pending; and
populating a secondary work queue with tasks having interprocess communication resources pending in response to said threshold has been exceeded, said secondary work queue being sorted according to a volume of resources pending for a task; and
executing tasks from said secondary work queue for a second time period before returning to the execution of tasks from said primary work queue, wherein said threshold, said first time period and said second time period are configurable at system initialization.
9. The method for scheduling tasks as recited in claim 8 , wherein said secondary work queue being further sorted according to a priority value associated with a task in said resources pending list.
10. The method for scheduling tasks as recited in claim 8 , wherein each of said resources has an upper limit.
11. The method for scheduling tasks as recited in claim 8 , wherein said resources are chosen from at least one of shared memory, semaphores, message queues, message pipes, sockets and signals.
12. The method for scheduling tasks as recited in claim 8 , wherein said secondary work queue is populated prior to execution from said secondary work queue.
13. The method for scheduling tasks as recited in claim 8 , wherein said primary work queue is executed according to at least one of preemptive priority scheduling, weighted round robin scheduling, and weighted deficit round robin scheduling.
14. The method for scheduling tasks as recited in claim 8 , wherein a given task that has received processing time in said primary work queue is not used to populate said secondary work queue.
15. A system, comprising:
means for populating a first work queue with a plurality of tasks;
means for processing said plurality of tasks from said first work queue for a first period of time;
means for maintaining a list of interprocess communication resources pending for each of said plurality of tasks during execution of tasks from said first work queue;
means for determining, after said first time period, whether said interprocess communication resources pending has exceeded a threshold based on said maintained list of interprocess communication resources;
means for populating a second work queue with tasks having interprocess communication resources pending in response to said threshold has been exceeded, wherein said second work queue is ordered according to a respective volume of interprocess communication resources pending for each of said plurality of tasks;
means for processing tasks from said second work queue for a second period of time before returning to the processing of tasks from said first work queue, wherein said threshold, said first period of time and said second period of time are configurable at system initialization; and
a processor.
16. The system as recited in claim 15 , wherein said means for populating said first work queue uses one of preemptive priority scheduling, weighted round robin scheduling, and weighted deficit round robin scheduling.
17. The system as recited in claim 15 , wherein said means for maintaining is implemented as hardware.
18. The system as recited in claim 15 , wherein a given task that has received processing time in said first work queue is not used to populate said second work queue.
19. The system as recited in claim 15 , wherein said interprocess communication resources are chosen from at least one of shared memory, semaphores, message queues, message pipes, sockets, and signals.
20. The system as recited in claim 15 , wherein said second work queue is populated prior to execution from said second work queue.