Generation of software thermal profiles executed on a set of processors using processor activity

• US 7,698,089 B2
• Filed: 11/29/2005
• Issued: 04/13/2010
• Est. Priority Date: 11/29/2005
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Johns, Charles Ray, Nutter, Mark Richard, Stafford, James Michael, Aguilar, Maximino Jr.
• Primary Examiner(s): Nghiem; Michael P.
• Assistant Examiner(s): KHUU, HIEN DIEU THI
• Application Number:
  US11/289,088
• Publication Number:
  US 20070124101A1
• Time in Patent Office: 1,596 Days
• Field of Search: 702/130, 702/186, 702/80, 702/182, 714/5, 714/10, 714/42, 714/47, 714/718, 374/100, 374/101
• US Class Current: 702/130
• CPC Class Codes: G01K 3/005 Circuits arrangements for indicating a predetermined temperature fire detection G08B17/00
• CPC Class Codes: G01K 7/015 using microstructures, e.g. made of silicon
• CPC Class Codes: G06F 1/206 comprising thermal management

Abstract

A computer implemented method, data processing system, computer usable code, and apparatus are provided for generation of software thermal profiles for applications executing on a set of processors. Sampling is performed of the hardware operations occurring in a set of processors during the execution of a set of workloads to create sampled information. A thermal index is then generated based on the sampled information.

Claims

1. A computer implemented method for generating software thermal profiles for applications executed on a set of processors, the computer implemented method comprising:

sampling hardware operations occurring in the set of processors during execution of a set of workloads for the applications by the set of processors to create sampled information, wherein sampling of sensors is performed in parallel based on an operation type information and an operation frequency information, in a non-simulated environment so that the sampled information is actual and non-predicted;

calculating a value of an inverse of cycles per instruction using the operation type information; and

generating a thermal index for the applications based on the sampled information and the value, wherein the thermal index is coupled to a particular hardware, workload and thermal environment.

2. The computer implemented method of claim 1, further comprising:

storing the sampled information in a data structure, wherein the thermal index is generated using the sampled information in the data structure and the value.

3. The computer implemented method of claim 1, wherein sampling the hardware operations is performed through use of instrumentation built into the set of processors.

4. The computer implemented method of claim 1, wherein the sampled information is at least one of instruction activity, cache activity, bus activity, or memory accesses of at least one of a set of execution units or a set of data flow units of the set of processors.

5. The computer implemented method of claim 4, wherein the thermal index is a weighted sum of the sampled information representing the intensity of at least one of the set of execution units or the set of data flow units.

6. The computer implemented method of claim 4, wherein the thermal index is a vector of values for each unit of at least one of the set of execution units of the set of processors or the set of data flow units of the set of processors.

7. The computer implemented method of claim 1, wherein the thermal index represents a thermal efficiency of the processor.

8. The computer implemented method of claim 1, wherein the hardware operations are sampled for a predetermined amount of time and the predetermined amount of time of a sample period is controlled by a configuration field.

9. The computer implemented method of claim 1, wherein the thermal index is proportional to the workload being executed.

10. The computer implemented method of claim 1, wherein the operation type information and the operation frequency information of the sampled information is analyzed to form analyzed information that with the thermal index are used by a compiler to optimize the thermal profile of an application.

11. The computer implemented method of claim 1, wherein the thermal index is a single number representing one of the value and a combination of the sampled information and the value.

12. A data processing system comprising:

a bus system;

a communications system connected to the bus system;

a memory connected to the bus system, wherein the memory includes a set of instructions; and

a processing unit connected to the bus system, wherein the processing unit executes the set of instructions to sample hardware operations occurring in a set of processors during execution of a set of workloads for applications by the set of processors to create sampled information, wherein sampling of sensors is performed in parallel based on an operation type information and an operation frequency information, in a non-simulated environment so that the sampled information is actual and non-predicted;

calculate a value of an inverse of cycles per instruction using the operation type information of the sampled information; and

generate a thermal index for the applications using the sampled information and the value wherein the thermal index is coupled to a particular hardware, workload and thermal environment.

13. The data processing system of claim 12, wherein the processing unit executes the set of instructions to store the sampled information in a data structure, wherein the thermal index is generated using the sampled information in the data structure and the value.

14. The data processing system of claim 12, wherein sampling the hardware operations is performed through use of instrumentation built into the set of processors.

15. The data processing system of claim 12, wherein the sampled information is at least one of instruction activity, cache activity, bus activity, or memory accesses of at least one of a set of execution units or a set of dataflow units of the set of processors.

16. The data processing system of claim 15, wherein the thermal index is a weighted sum of the sampled information representing the intensity of at least one of the set of execution units or the set of data flow units.

17. The data processing system of claim 15, wherein the thermal index is a vector of values for each unit of at least one of the set of execution units of the set of processors or the set of data flow units of the set of processors.

18. The data processing system of claim 12, wherein the thermal index represents a thermal efficiency of the processor.

19. The data processing system of claim 12, wherein the hardware operations are sampled for a predetermined amount of time and the predetermined amount time of a sample period is controlled by a configuration field.

20. The data processing system of claim 12, wherein the thermal index is proportional to the workload being executed.

21. The data processing system of claim 12, wherein the operation type information and the operation frequency information of the sampled information is analyzed to form analyzed information that with the thermal index are used by a compiler to optimize the thermal profile of an application.

22. The data processing system of claim 12, wherein the thermal index is a single number representing one of the value and a combination of the sampled information and the value.

23. A computer program product for generating software thermal profiles for applications executed on a set of processors comprising:

a computer usable memory element including computer usable program code stored thereon, the computer usable program code including;

computer usable program code for sampling hardware operations occurring in the set of processors during execution of a set of workloads for the applications by the set of processors to create sampled information, wherein sampling of sensors is performed in parallel based on an operation type information and an operation frequency information in a non-simulated environment so that the sampled information is actual and non-predicted;

computer usable program code for calculating a value of an inverse of cycles per instruction using the operation type information; and

computer usable program code for generating a thermal index for the applications based on the sampled information and the value, wherein the thermal index is coupled to a particular hardware, workload and thermal environment.

24. The computer program product of claim 23, further including:

computer usable program code for storing the sampled information in a data structure, wherein the thermal index is generated using the sampled information in the data structure and the value.

25. The computer program product of claim 23, wherein sampling the hardware operations is performed through use of instrumentation built into the set of processors.

26. The computer program product of claim 23, wherein the sampled information is at least one of instruction activity, cache activity, bus activity, or memory accesses of at least one of a set of execution units or a set of dataflow units of the set of processors.

27. The computer program product of claim 26, wherein the thermal index is a weighted sum of the sampled information representing the intensity of at least one of the set of execution units or the set of data flow units.

28. The computer program product of claim 26, wherein the thermal index is a vector of values for each unit of at least one of the set of execution units of the set of processors or the set of data flow units of the set of processors.

29. The computer program product of claim 23, wherein the thermal index represents a thermal efficiency of the processor.

30. The computer program product of claim 23, wherein the hardware operations are sampled for a predetermined amount of time and the predetermined amount time of a sample period is controlled by a configuration field.

31. The computer program product of claim 23, wherein the thermal index is proportional to the workload being executed.

32. The computer program product of claim 23, wherein the operation type information and the operation frequency information of the sampled information is analyzed to form analyzed information that with the thermal index are used by a compiler to optimize the thermal profile of an application.

33. The computer program product of claim 23, wherein the thermal index is a single number representing one of the value and a combination of the sampled information and the value.

34. An apparatus for generating software thermal profiles for applications executed on a set of processors, the apparatus comprising:

instrumentation for sampling hardware operations occurring in the set of processors during execution of a set of workloads for the applications by the set of processors to create sampled information, wherein sampling of sensors is performed in parallel based on an operation type information and an operation frequency information, for a predetermined amount of time of a sample period as controlled by a configuration field in a non-simulated environment so that the sampled information is actual and non-predicted; and

a generator for generating a value of an inverse of cycles per instruction using the operation type information; and

a thermal index for the applications based on the sampled information and the value, wherein the thermal index is coupled to a particular hardware, workload and thermal environment.

35. The apparatus of claim 34, further comprising:

storage for storing the sampled information in a data structure, wherein the thermal index is generated using the sampled information in the data structure and the value.