METHODS, SYSTEMS, ARTICLES OF MANUFACTURE, AND APPARATUS TO OPTIMIZE THREAD SCHEDULING

• US 20200133725A1
• Filed: 12/27/2019
• Published: 04/30/2020
• Est. Priority Date: 12/02/2019
• Status: Active Grant

• Current Assignee: Intel Corporation
• Original Assignee: Intel Corporation
• Inventors: Thomas, Renji, Binns, Chris, Mercati, Pietro, Zhu, Jianfang, Wadaa, Ashraf H., Kishinevsky, Michael, Shams, Ahmed
• Granted Patent: US 11,782,755 B2
• Time in Patent Office: Days
• CPC Class Codes: G06F 1/3228 Monitoring task completion, e.g. by use of idle timers, stop commands or wait commands
• CPC Class Codes: G06F 1/329 by task scheduling
• CPC Class Codes: G06F 11/0715 in a system implementing multitasking multitasking per se G06F9/46
• CPC Class Codes: G06F 11/0751 Error or fault detection not based on redundancy power supply failures G06F1/30; network fault management H04L41/06
• CPC Class Codes: G06F 11/0793 Remedial or corrective actions recovery from an exception in an instruction pipeline G06F9/3861; by retry G06F11/1402; for recovering from a failure of a protocol instance or entity H04L69/40
• CPC Class Codes: G06F 11/3017 where the computing system is implementing multitasking multiprogramming arrangements G06F9/46; allocation of resources G06F9/50
• CPC Class Codes: G06F 11/3428 Benchmarking
• CPC Class Codes: G06F 11/3433 for load management allocation of a server based on load conditions G06F9/505; load rebalancing G06F9/5083; redistributing the load in a network by a load balancer H04L67/1029
• CPC Class Codes: G06F 2119/06 Power analysis or power optimisation
• CPC Class Codes: G06F 2209/501 Performance criteria
• CPC Class Codes: G06F 2209/5018 Thread allocation
• CPC Class Codes: G06F 2209/508 Monitor
• CPC Class Codes: G06F 30/20 Design optimisation, verification or simulation optimisation, verification or simulation of circuit designs G06F30/30
• CPC Class Codes: G06F 30/27 using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
• CPC Class Codes: G06F 9/3836 Instruction issuing, e.g. dynamic instruction scheduling or out of order instruction execution
• CPC Class Codes: G06F 9/4881 Scheduling strategies for dispatcher, e.g. round robin, multi-level priority queues
• CPC Class Codes: G06F 9/4893 taking into account power or heat criteria power management in computers in general G06F1/3203; thermal management in computers in general G06F1/206
• CPC Class Codes: G06F 9/5027 the resource being a machine, e.g. CPUs, Servers, Terminals
• CPC Class Codes: G06F 9/5094 where the allocation takes into account power or heat criteria power management in computers in general G06F1/3203; thermal management in computers in general G06F1/206
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 7/01 : Probabilistic graphical models, e.g. probabilistic networks
• CPC Class Codes: Y02D 10/00 : Energy efficient computing, e.g. low power processors, power management or thermal management

Abstract

An apparatus comprising: a model to generate adjusted tuning parameters of a thread scheduling policy based on a tradeoff indication value of a target system; and a workload monitor to: execute a workload based on the thread scheduling policy; obtain a performance score and a power score from the target system based on execution of the workload, the performance score and the power score corresponding to a tradeoff indication value; compare the tradeoff indication value to a criterion; and based on the comparison, initiate the model to re-adjust the adjusted tuning parameters.

Claims

1. An apparatus comprising:

a model to generate adjusted tuning parameters of a thread scheduling policy based on a tradeoff indication value of a target system; and

a workload monitor to;

execute a workload based on the thread scheduling policy;

obtain a performance score and a power score from the target system based on execution of the workload, the performance score and the power score corresponding to a tradeoff indication value;

compare the tradeoff indication value to a criterion; and

based on the comparison, initiate the model to re-adjust the adjusted tuning parameters.

2. The apparatus of claim 1, further including a performance evaluation controller to determine the performance score of the target system during multiple iterations of evaluating the workload execution to re-adjust the adjusted tuning parameters based on the performance score.

3. The apparatus of claim 1, further including a power evaluation controller to determine the power score of the target system during multiple iterations of evaluating the workload execution to re-adjust the adjusted tuning parameters based on the power score.

4. The apparatus of claim 1, further including an error checking and recovery controller to:

detect an unexpected state of the target system; and

revert the target system to a last known good state to enable multiple iterations of evaluating the thread scheduling policy to continue re-adjusting the adjusted tuning parameters.

5. The apparatus of claim 1, wherein the workload monitor is to determine a point of the adjusted tuning parameters of the model at which the adjusted tuning parameters are maximized.

6. The apparatus of claim 1, wherein the workload monitor is to obtain a baseline performance score corresponding to a minimum performance of the target system and a baseline power score corresponding to a minimum power score of the target system.

7. The apparatus of claim 6, wherein the model is to determine initial tuning parameters based on the baseline performance score and the baseline power score of the target system, the model to generate the adjusted tuning parameters based on the initial tuning parameters to configure the thread scheduling policy for a target optimization.

8. A non-transitory computer readable storage medium comprising instructions that, when executed, cause a machine to at least:

generate adjusted tuning parameters of a thread scheduling policy based on a tradeoff indication value of a target system; and

execute a workload based on the thread scheduling policy;

obtain a performance score and a power score from the target system based on execution of the workload, the performance score and the power score corresponding to a tradeoff indication value;

compare the tradeoff indication value to a criterion; and

based on the comparison, initiate a model to re-adjust the adjusted tuning parameters.

9. The non-transitory computer readable storage medium of claim 8, wherein the instructions, when executed, cause the machine to determine the performance score of the target system during multiple iterations of evaluating the workload execution to re-adjust the adjusted tuning parameters based on the performance score.

10. The non-transitory computer readable storage medium of claim 8, wherein the instructions, when executed, cause the machine to determine the power score of the target system during multiple iterations of evaluating the workload execution to re-adjust the adjusted tuning parameters based on the power score.

11. The non-transitory computer readable storage medium of claim 8, wherein the instructions, when executed, cause the machine to:

detect an unexpected state of the target system; and

revert the target system to a last known good state to enable multiple iterations of evaluating the thread scheduling policy to continue re-adjusting the adjusted tuning parameters.

12. The non-transitory computer readable storage medium of claim 8, wherein the instructions, when executed, cause the machine to determine a point of the adjusted tuning parameters of the model at which the adjusted tuning parameters are maximized.

13. The non-transitory computer readable storage medium of claim 8, wherein the instructions, when executed, cause the machine to obtain a baseline performance score corresponding to a minimum performance of the target system and a baseline power score corresponding to a minimum power score of the target system.

14. The non-transitory computer readable storage medium of claim 13, wherein the instructions, when executed, cause the machine to determine initial tuning parameters based on the baseline performance score and the baseline power score of the target system, the instructions to generate the adjusted tuning parameters based on the initial tuning parameters to configure the thread scheduling policy for a target optimization.

15. An apparatus comprising:

means for generating adjusted tuning parameters of a thread scheduling policy based on a tradeoff indication value of a target system;

means for monitoring to;

execute a workload based on the thread scheduling policy;

obtain a performance score and a power score from the target system based on execution of the workload, the performance score and the power score corresponding to a tradeoff indication value;

compare the tradeoff indication value to a criterion; and

based on the comparison, initiate a model to re-adjust the adjusted tuning parameters.

16. The apparatus of claim 15, further including means for determining the performance score of the target system during multiple iterations of evaluating the workload execution to re-adjust the adjusted tuning parameters based on the performance score.

17. The apparatus of claim 15, further including means for determining the power score of the target system during multiple iterations of evaluating the workload execution to re-adjust the adjusted tuning parameters based on the power score.

18. The apparatus of claim 15, further including a means for checking to:

detect an unexpected state of the target system; and

revert the target system to a last known good state to enable multiple iterations of evaluating the thread scheduling policy to continue re-adjusting the adjusted tuning parameters.

19. The apparatus of claim 15, wherein the means for monitoring is configured to determine a point of the adjusted tuning parameters of the model at which the adjusted tuning parameters are maximized.

20. The apparatus of claim 15, wherein the means for monitoring is configured to obtain a baseline performance score corresponding to a minimum performance of the target system and a baseline power score corresponding to a minimum power score of the target system.

21. The apparatus of claim 20, wherein the means for monitoring is configured to determine initial tuning parameters based on the baseline performance score and the baseline power score of the target system, the instructions to generate the adjusted tuning parameters based on the initial tuning parameters to configure the thread scheduling policy for a target optimization.

22. A method comprising:

generating adjusted tuning parameters of a thread scheduling policy based on a tradeoff indication value of a target system; and

executing a workload based on the thread scheduling policy;

obtaining a performance score and a power score from the target system based on execution of the workload, the performance score and the power score corresponding to a tradeoff indication value;

comparing the tradeoff indication value to a criterion; and

based on the comparison, initiating a model to re-adjust the adjusted tuning parameters.

23. The method of claim 22, further including determining the performance score of the target system during multiple iterations of evaluating the workload execution to re-adjust the adjusted tuning parameters based on the performance score.

24. The method of claim 22, further including determining the power score of the target system during multiple iterations of evaluating the workload execution to re-adjust the adjusted tuning parameters based on the power score.

25. The method of claim 22, further including:

detecting an unexpected state of the target system; and

reverting the target system to a last known good state to enable multiple iterations of evaluating the thread scheduling policy to continue re-adjusting the adjusted tuning parameters.

26-28. -28. (canceled)