Method and apparatus for software-assisted thermal management for electronic systems

• US 6,948,082 B2
• Filed: 05/17/2002
• Issued: 09/20/2005
• Est. Priority Date: 05/17/2002
• Status: Active Grant

• Current Assignee: Focus Global Solutions LLC (Jeffrey M. Gross)
• Original Assignee: International Business Machines Corporation
• Inventors: Gschwind, Michael Karl, Salapura, Valentina
• Primary Examiner(s): Browne, Lynne H.
• Assistant Examiner(s): Trujillo, James K.
• Application Number:
  US10/150,270
• Publication Number:
  US 20030217297A1
• Time in Patent Office: 1,222 Days
• Field of Search: 713/300, 713/320, 713/322, 713/323, 713/324
• US Class Current: 713/320
• CPC Class Codes: G06F 1/206 comprising thermal management
• CPC Class Codes: Y02D 10/00 Energy efficient computing, e.g. low power processors, power management or thermal management

Abstract

In a computer system, a device for measuring power dissipation (e.g., using on-die thermal sensors) is linked to both a hardware-based thermal management solution and with a means for causing a notification event to software, so that, initially, the operating system software and/or the application software modifies its behavior in response to the notification event to reduce overall system power dissipation and the hardware-based thermal management solution is only triggered if the software solution is not effective; with both operating system and application software resuming higher-performance algorithms when power dissipation is no longer critical.

Claims

1. A computer system comprising an integrated circuit CPU, and at least one software program, in which said CPU includes sensor means responsive to a thermal property of said integrated circuit for asserting a first notification event;

and said software program includes instruction means responsive to said first notification event for substituting at least one corresponding power-saving algorithm for an algorithm executed by said software program to reduce the thermal load imposed by said software program on said CPU.

2. A computer system according to claim 1, further comprising means within said CPU for asserting a second notification event when said thermal property exceeds a second limit.

3. A computer system according to claim 2, in which said second limit is greater than said first limit;

and

further comprising means for providing a hardware-based thermal management solution in response to said second notification event.

4. A computer system according to claim 1, in which said instruction means are included within operating system software.

5. A computer system according to claim 4, in which said operating system software modifies the execution of at least one operating application program in response to said first notification event.

6. A computer system according to claim 4, in which said power-saving algorithm affects a scheduling of at least one resource on a unit of said integrated circuit CPU to reduce said thermal load.

7. A computer system according to claim 4, in which said power-saving algorithm affects a region on said integrated circuit CPU in which said power-saving algorithm runs.

8. A computer system according to claim 1, in which said instruction means are included within application software.

9. A method of operating a computer system comprising an integrated circuit CPU and at least one software program, in which said CPU includes sensor means responsive to a thermal property of said integrated circuit for asserting a first notification event, comprising the steps of:

sensing a parameter representing said first notification event;

substituting at least one corresponding power-saving algorithm executed by said software program for an algorithm that reduces the thermal load imposed by said software program on said CPU in response to said assertion of a first notification event; and

restoring said original step in response to the withdrawal of said assertion of a first notification event.

10. A method according to claim 9, in which said substitute step suspends the operation of at least one application program.

11. A method according to claim 9, in which said substitute step substitutes at least one power-saving parameter in an algorithm for a corresponding parameter.

12. A method according to claim 9, in which said substitute step suspends the operation of at least one algorithm in an application program.

13. A method according to claim 9, in which said assertion of a first notification event is in the form of setting a value of said parameter and further comprising a step of comparing said parameter with a first reference value;

and modifying said method executed by said software program by the substitution for said original step of a first substitute step that reduces the thermal load imposed by said software program on said CPU when said value of said parameter exceeds said first value.

14. A method according to claim 13, further comprising a step of comparing said parameter with a first restoring reference value;

and

modifying said method executed by said software program by the replacement of said original step when said value of said parameter is less than said first restoring reference value.

15. A method according to claim 13, further comprising a step of comparing said parameter with a second reference value;

and

modifying said method executed by said software program by the substitution for said first substitute step of a second substitute step that further reduces said thermal load imposed by said software program on said CPU when said value of said parameter passes said second reference value.

16. A method according to claim 15, further comprising a step of comparing said parameter with a second restoring reference value;

and

modifying said method executed by said software program by the replacement of said second substitute step with said first substitute step when said value of said parameter is less than said second restoring reference value.

17. A method of operating a computer system comprising an integrated circuit CPU and at least one software program, in which said CPU includes sensor means responsive to a thermal property of said integrated circuit for asserting a first notification event, comprising the steps of:

sensing a parameter representing said first notification event;

modifying a method executed by said software program by the substitution for an original step of at least one substitute step that reduces the thermal load imposed by said software program on said CPU in response to said assertion of a first notification event; and

restoring said original step in response to the withdrawal of said assertion of a first notification event wherein said original step is restored automatically after the passage of a time-out interval.

18. An article of manufacture for use in a computer system having an integrated circuit CPU having sensor means responsive to a thermal property of said integrated circuit for asserting a first notification event, comprising a software program storage medium readable by a computer and embodying instructions executable by the computer for performing a method comprising the steps of:

sensing a parameter representing said first notification event;

modifying said method executed by said software program by the substitution for an original step of at least one substitute step that reduces the thermal load imposed by said software program on said CPU in response to said assertion of a first notification event; and

restoring said original step in response to the withdrawal of said assertion of a first notification event wherein modifying said method comprises substituting at least one corresponding power-saving algorithm for an algorithm.

19. An article of manufacture according to claim 18, in which said substitute step suspends the operation of at least one application program.

20. An article of manufacture according to claim 18, in which said substitute step substitutes at least one power-saving parameter in an algorithm for a corresponding parameter.

21. An article of manufacture according to claim 18, in which said substitute step suspends the operation of at least one algorithm in an application program.

22. An article of manufacture according to claim 18, in which said assertion of a first notification event is in the form of setting a value of a parameter and further comprising a step of comparing said parameter with a first reference value;

and

modifying said method executed by said software program by the substitution for said original step of a first substitute step that reduces the thermal load imposed by said software program on said CPU when said value of said parameter exceeds said first reference value.

23. An article of manufacture according to claim 22, further comprising a step of comparing said parameter with a first restoring reference value;

and

modifying said method executed by said software program by the replacement of said original step when said value of said parameter is less than said first restoring reference value.

24. An article of manufacture according to claim 22, further comprising a step of comparing said parameter with a second reference value;

and

modifying said method executed by said software program by the substitution for said first substitute step of a second substitute step that further reduces said thermal load imposed by said software program on said CPU when said value of said parameter passes said second reference value.

25. An article of manufacture according to claim 24, further comprising a step of comparing said parameter with a second restoring reference value;

and

modifying said method executed by said software program by the replacement of said second substitute step with said first substitute step when said value of said parameter is less than said second restoring reference value.

26. An article of manufacture for use in a computer system having an integrated circuit CPU having sensor means responsive to a thermal property of said integrated circuit for asserting a first notification event, comprising a software program storage medium readable by a computer and embodying instructions executable by the computer for performing a method comprising the steps of:

sensing a parameter representing said first notification event;

modifying said method executed by said software program by the substitution for an original step of at least one substitute step that reduces the thermal load imposed by said software program on said CPU in response to said assertion of a first notification event; and

restoring said original step in response to the withdrawal of said assertion of a first notification event wherein said original step is restored automatically after the passage of a time-out interval.