Inter-thread trace alignment method and system for a multi-threaded processor

• US 8,484,516 B2
• Filed: 04/11/2007
• Issued: 07/09/2013
• Est. Priority Date: 04/11/2007
• Status: Active Grant

• Current Assignee: Qualcomm, Inc.
• Original Assignee: Qualcomm, Inc.
• Inventors: Giannini, Louis Achille, Anderson, William, Chen, Xufeng
• Primary Examiner(s): Beausoliel, Robert
• Assistant Examiner(s): Lottich, Joshua P
• Application Number:
  US11/734,199
• Publication Number:
  US 20080256396A1
• Time in Patent Office: 2,281 Days
• Field of Search: 714/12, 714/20, 714/38, 714/45, 717/128
• US Class Current: 714/45
• CPC Class Codes: G06F 11/3636 by tracing the execution of the program

Abstract

Techniques for the design and use of a digital signal processor, including (but not limited to) for processing transmissions in a communications (e.g., CDMA) system. Inter-thread trace alignment with execution trace processing includes recording timing data relating to a common predetermined event. Such an event may be the number of cycles since a last thread initiated execution tracing or the number of cycles since all threads terminated execution tracing. The number of cycles at which a thread initiates execution tracing is referenced to the common predetermined event for maintaining the timing of execution tracing. The data relating to the common predetermined event is then updated to associate with the time at which the thread initiated execution tracing. The result is to permit aligning the timing data associated with all threads. Interrelated records permit reconstructing interdependent execution tracing information for threads operating in the multi-threaded processor, as well as synchronizing timing data for all operating threads.

Claims

1. A method for inter-thread trace timing alignment among threads of a multi-threaded processor during an execution tracing process, comprising:

recording timing data relating to a common predetermined event, said common predetermined event being commonly referenceable by all operating threads of said multi-threaded processor during core processor execution tracing;

referencing a time at which a thread initiates execution tracing to said common predetermined event for maintaining the timing of execution tracing for said thread relative to said common predetermined event; and

updating said common predetermined event to associate with said time at which said thread initiated execution tracing, thereby aligning said time at which said thread initiated execution tracing to timing data associated with all other threads of said multi-threaded processor for which execution tracing may be occurring.

2. The method of claim 1, further comprising relating said common predetermined event to a time at which a last thread turned on execution tracing.

3. The method of claim 1, further comprising relating said common predetermined event to the number of multi-threaded processor cycles since all threads turned off execution tracing.

4. The method of claim 1, further comprising recording said common predetermined event in a plurality of data packets.

5. The method of claim 1, further comprising reconstructing interrelated execution tracing information for threads operating in said multi-threaded processor using a plurality of data packets associated with said common predetermined event.

6. The method of claim 1, further comprising synchronizing said timing data relating to said common predetermined event.

7. The method of claim 1, further comprising generating a global count of cycles since the occurrence of said common predetermined event.

8. The method of claim 1, further comprising re-establishing inter-thread timing data among all threads performing execution tracing using a global count of cycles since said common predetermined event.

9. The method of claim 1, further comprising determining a data loss and, in response to said data loss, re-establishing inter-thread timing data among all threads performing execution tracing using a global count of cycles since said common predetermined event, in the event of a data loss.

10. The method of claim 1, further comprising generating a synchronization packet for all threads in the event of a global counter cycling through a zero value.

11. A hardware digital signal processor debugging system for operation in association with a digital signal processor and including the ability to interrelate timing data relating to threads of a multi-threaded processor during multi-threaded software execution flow, comprising:

a plurality of offset fields for recording timing data relating to a common predetermined event, said common predetermined event being commonly referenceable by all operating threads of said multi-threaded processor during core processor execution tracing;

time referencing instruction packets for referencing a time at which a thread initiates execution tracing to said common predetermined event for maintaining the timing of execution tracing for said thread relative to said common predetermined event; and

time reference updating packets for updating said common predetermined event to associate with said time at which said thread initiated execution tracing, thereby aligning said time at which said thread initiated execution tracing to timing data associated with all other threads of said multi-threaded processor for which execution tracing may be occurring.

12. The digital signal processor debugging system of claim 11, further comprising circuitry and instructions for relating said common predetermined event to the time at which a last thread turned on execution tracing.

13. The digital signal processor debugging system of claim 11, further comprising circuitry and instructions for relating said common predetermined event to the number of multi-threaded processor cycles since all threads turned off execution tracing.

14. The digital signal processor debugging system of claim 11, further comprising circuitry and instructions for recording said common predetermined event in a plurality of data packets.

15. The digital signal processor debugging system of claim 11, further comprising circuitry and instructions for reconstructing interrelated execution tracing information for threads operating in said multi-threaded processor using a plurality of data packets associated with said common predetermined event.

16. The digital signal processor debugging system of claim 11, further comprising circuitry and instructions for synchronizing said timing data relating to said common predetermined event.

17. The digital signal processor debugging system of claim 11, further comprising circuitry and instructions for generating a global count of cycles since the occurrence of said common predetermined event.

18. The digital signal processor debugging system of claim 11, further comprising thread interrelating instructions for re-establishing inter-thread timing data among all threads performing execution tracing using a global count of cycles since said common predetermined event.

19. The digital signal processor debugging system of claim 11, further comprising execution tracing instruction for determining a data loss, in response to said data loss, re-establishing inter-thread timing data among all threads performing execution tracing using a global count of cycles since said common predetermined event.

20. The digital signal processor debugging system of claim 11, further comprising synchronization instructions for generating a synchronization packet for all threads in the event of a global counter cycling through a zero value.

21. A multi-threaded digital signal processor for operation in support of a personal electronics device, the multi-threaded digital signal processor comprising debugging means for performing an execution tracing process and, in relation thereto, aligning inter-thread trace timing among threads of said multi-threaded processor, comprising:

means for recording timing data relating to a common predetermined event, said common predetermined event being commonly referenceable by all operating threads of said multi-threaded processor during core processor execution tracing;

means for referencing a time at which a thread initiates execution tracing to said common predetermined event for maintaining the timing of execution tracing for said thread relative to said common predetermined event; and

means for updating said common predetermined event to associate with said time at which said thread initiated execution tracing, thereby aligning said time at which said thread initiated execution tracing to timing data associated with all other threads of said multi-threaded processor for which execution tracing may be occurring.

22. The digital signal processor system of claim 21, further comprising means for relating said common predetermined event to the time at which a last thread turned on execution tracing.

23. The digital signal processor system of claim 21, further comprising means for relating said common predetermined event to the number of multi-threaded processor cycles since all threads turned off execution tracing.

24. The digital signal processor system of claim 21, further comprising means for recording said common predetermined event in a plurality of data packets.

25. The digital signal processor system of claim 21, further comprising means for reconstructing interrelated execution tracing information for threads operating in said multi-threaded processor using a plurality of data packets associated with said common predetermined event.

26. The digital signal processor system of claim 21, further comprising means for synchronizing said timing data relating to said common predetermined event.

27. The digital signal processor system of claim 21, further comprising means for generating a global count of cycles since the occurrence of said common predetermined event.

28. The digital signal processor system of claim 21, further comprising means for re-establishing inter-thread timing data among all threads performing execution tracing using a global count of cycles since said common predetermined event.

29. The digital signal processor system of claim 21, further comprising means for determining a data loss and, in response to said data loss, re-establishing inter-thread timing data among all threads performing execution tracing using a global count of cycles since said common predetermined event.

30. The digital signal processor system of claim 21, further comprising means for generating a synchronization packet for all threads in the event of a global counter cycling through a zero value.

31. A non-transitory computer usable medium having computer readable program code means embodied therein for debugging a multi-threaded digital signal processor including performing an execution tracing process and, in relation thereto, aligning inter-thread trace timing among threads of said multi-threaded processor, the computer usable medium comprising:

computer readable program code means for recording timing data relating to a common predetermined event, said common predetermined event being commonly referenceable by all operating threads of said multi-threaded processor during core processor execution tracing;

computer readable program code means for referencing a time at which a thread initiates execution tracing to said common predetermined event for maintaining the timing of execution tracing for said thread relative to said common predetermined event; and

computer readable program code means for updating said common predetermined event to associate with said time at which said thread initiated execution tracing, thereby aligning said time at which said thread initiated execution tracing to timing data associated with all other threads of said multi-threaded processor for which execution tracing may be occurring.

32. The computer usable medium of claim 31, further comprising computer readable program code means for relating said common predetermined event to the time at which a last thread turned on execution tracing.

33. The computer usable medium of claim 31, further comprising computer readable program code means for relating said common predetermined event to the number of multi-threaded processor cycles since all threads turned off execution tracing.

34. The computer usable medium of claim 31, further comprising computer readable program code means for recording said common predetermined event in a plurality of data packets.

35. The computer usable medium of claim 31, further comprising computer readable program code means for reconstructing interrelated execution tracing information for threads operating in said multi-threaded processor using a plurality of data packets associated with said common predetermined event.