Energy aware processing load distribution system and method

Patent metadata

Field	Value
Publication number	US-11809254-B2
Application number	US-202117543122-A
Country	US
Kind code	B2
Filing date	Dec 6, 2021
Priority date	Jul 21, 2009
Publication date	Nov 7, 2023
Grant date	Nov 7, 2023

Abstract

A method for controlling a data center, comprising a plurality of server systems, each associated with a cooling system and a thermal constraint, comprising: a concurrent physical condition of a first server system; predicting a future physical condition based on a set of future states of the first server system; dynamically controlling the cooling system in response to at least the input and the predicted future physical condition, to selectively cool the first server system sufficient to meet the predetermined thermal constraint; and controlling an allocation of tasks between the plurality of server systems to selectively load the first server system within the predetermined thermal constraint and selectively idle a second server system, wherein the idle second server system can be recruited to accept tasks when allocated to it, and wherein the cooling system associated with the idle second server system is selectively operated in a low power consumption state.

First claim

What is claimed is:

1. A scheduler for allocating tasks for performance by a respective processor of a plurality of available processors, comprising:
at least one table representing energy/performance characteristics, thermal characteristics, and utilization statistics associated with each respective available processor in each of a plurality of different performance settings;
a memory configured to store respective quality of service requirements for different types of tasks;
an input configured to receive thermal information;
at least one input configured to receive data representing a characteristic of an ordered queue of uncompleted tasks for each respective available processor, dynamically dependent on the uncompleted tasks in the respective ordered queue;
at least one automated processor configured to:
determine a type of a task,
control a performance setting of the plurality of available processors,
selectively assign a placement of the task within the ordered queue of a respective available processor of the plurality of available processors in dependence on:
the received thermal information,
the determined type of the tasks within the ordered queue of the respective available processor,
the thermal information,
the quality of service requirements for the type of task, and
the at least one table representing the energy/performance characteristics, the thermal characteristics, and the utilization statistics associated with each respective available processor; and

an output configured to reflect the assigned placement of the task in the ordered queue for the task.

2. The scheduler according to claim 1 , wherein the characteristic of the ordered queue of uncompleted tasks comprises a size of the respective ordered queue.

3. The scheduler according to claim 1 , wherein the at least one automated processor is further configured to proactively schedule a cooling system for a respective available processor in dependence on at least a content of the ordered queue of the respective available processor.

4. The scheduler according to claim 1 , wherein the at least one automated processor is further configured to make a previously unavailable processor available, and make an respective available processor unavailable.

5. The scheduler according to claim 1 , wherein the at least one automated processor is further configured to determine a respective type and an energy requirement for a task based on a past energy requirement for at least one previously processed task.

6. The scheduler according to claim 1 , wherein the plurality of respective available processors comprise respective processing cores of a multicore processor.

7. The scheduler according to claim 1 , wherein the at least one automated processor is further configured to produce a signal for controlling a voltage and frequency setting state associated with each respective available processor.

8. A scheduler for allocating tasks for execution by a respective processor of a plurality of available processors, comprising:
at least one table representing energy consumption characteristics, performance characteristics, thermal characteristics, and utilization statistics associated with each respective available processor dependent on a performance setting;
a memory configured to store respective quality of service requirements for different types of tasks in a memory, wherein the quality of service requirements comprise a maximum permissible latency of execution;
an input configured to receive thermal information relating to the plurality of available processors;
an ordered queue of uncompleted tasks for each respective available processor, having an associated dynamically changing characteristic dependent on the uncompleted tasks in the respective ordered queue; and
at least one automated processor configured to determine a type of a respective task, the performance setting for the plurality of available processors, and selectively assign a placement of the respective task in the ordered queue of a respective available processor of the plurality of available processors in dependence on the received thermal information, the determined types of the tasks in the ordered queue of the respective processor, the dynamically changing characteristic, the quality of service requirements for the type of task, and the at least one table.

9. The scheduler according to claim 8 , wherein the dynamically changing characteristic of the ordered queue of uncompleted tasks comprises a size of the respective ordered queue.

10. The scheduler according to claim 8 , wherein the at least one automated processor is further configured to proactively schedule a cooling system for a respective available processor in dependence on at least a content of the ordered queue of the respective available processor.

11. The scheduler according to claim 8 , wherein the at least one automated processor is further configured to make an additional processor available, having a respective ordered queue, and to make a respective available processor unavailable and to place tasks in the respective ordered queue of the unavailable processor within the ordered queues of remaining available processors.

12. The scheduler according to claim 8 , wherein the at least one automated processor is further configured to determine the type and an energy requirement for the respective task based on a past energy requirement for at least one previously processed task.

13. The scheduler according to claim 8 , wherein the plurality of available processors comprise respective processors of a multicore processor.

14. The scheduler according to claim 8 , wherein the at least one automated processor is further configured to control a voltage and frequency setting state associated with each respective available processor.

15. The scheduler according to claim 8 , wherein the at least one automated processor is further configured to allocate an order of tasks within a respective ordered queue for the respective available processor to optimize a predicted energy efficiency of plurality of available processors while ensuring that the respective quality of service requirement for each task is fulfilled and a thermal constraint on the respective available processor is not exceeded.

16. The scheduler according to claim 8 , wherein the at least one automated processor is further configured to reallocate at least one task from an ordered queue for a first available processor to an ordered queue for a second available processor.

17. The scheduler according to claim 8 , wherein a sequencing of tasks within a respective queue is further responsive to a latency of completion of at least one task by the respective available processor.

18. The scheduler according to claim 8 , wherein the at least one automated processor is configured to allocate the task to a respective ordered queue to selectively unbalance loads on the plurality of available processors for increased cooling efficiency.

19. A scheduler, comprising:
a first input port configured to receive a task;
a second input port configured to receive thermal information relating to a plurality of respective processors;
a table representing energy-performance characteristics, thermal characteristics, and utilization statistics associated with performance settings for each of the plurality of respective processors;
a first memory configured to store quality of service requirements for different types of tasks;
a second memory configured to store data representing a characteristic of an ordere

Assignees

• Univ New York State Res Found
• The Res Foundation For Suny

Inventors

• Ghose Kanad

Classifications

• A61K36/185

  Magnoliopsida (dicotyledons) · CPC title

• Y02D10/00

  Energy efficient computing, e.g. low power processors, power management or thermal management · CPC title

• G06F1/3209 Primary

  Monitoring remote activity, e.g. over telephone lines or network connections · CPC title

• G06F1/206 Primary

  comprising thermal management · CPC title

• B03B1/00

  Conditioning for facilitating separation by altering physical properties of the matter to be treated · CPC title