System and method for identifying contention of shared resources in a runtime system

• US 9,280,444 B2
• Filed: 10/11/2010
• Issued: 03/08/2016
• Est. Priority Date: 10/11/2010
• Status: Active Grant

• Current Assignee: SAP SE
• Original Assignee: SAP SE
• Inventors: Scheerer, Johannes, Schmelter, Ralf, Wintergerst, Michael, Schreiber, Steffen, Mostowoj, Dietrich
• Primary Examiner(s): Kim, Dong
• Assistant Examiner(s): MILLS, PAUL V
• Application Number:
  US12/901,907
• Publication Number:
  US 20120089991A1
• Time in Patent Office: 1,975 Days
• US Class Current: 1/1
• CPC Class Codes: G06F 11/323 Visualisation of programs or trace data
• CPC Class Codes: G06F 11/3632 of specific synchronisation aspects
• CPC Class Codes: G06F 11/3664 Environments for testing or debugging software
• CPC Class Codes: G06F 12/0269 Incremental or concurrent garbage collection, e.g. in real-time systems G06F12/0261 takes precedence
• CPC Class Codes: G06F 9/52 Program synchronisation; Mutual exclusion, e.g. by means of semaphores

Abstract

A system and computer-implemented method for determining a runtime of a thread of an application. Synchronization events for a first thread of an application executing on the computer system are received, the synchronization events including at least a first synchronization event and a second synchronization event for the first thread. A first difference between a synchronization event timestamp of the first synchronization event and the synchronization event timestamp of the second synchronization event is calculated. A second difference between an accumulated timestamp of the first synchronization event and the accumulated timestamp of the second synchronization event is calculated. A runtime of the first thread of the application is calculated as a difference between the first difference and the second difference.

Claims

1. A computer-implemented method of determining a runtime of a thread of an application, the method comprising:

receiving synchronization events for a first thread of an application executing on a computer system, the synchronization events including at least a first synchronization event and a second synchronization event for the first thread, each synchronization event includes a synchronization event timestamp corresponding to a time at which the synchronization event occurred and an accumulated timestamp corresponding to a time during which a runtime system of the computer system has been executing runtime system operations that include generating at least one stack trace of the first thread;

calculating a first difference between the synchronization event timestamp of the first synchronization event and the synchronization event timestamp of the second synchronization event;

calculating a second difference between the accumulated timestamp of the first synchronization event and the accumulated timestamp of the second synchronization event;

calculating, by at least one processor of a computer system, a runtime of the first thread of the application as a difference between the first difference and the second difference;

providing a first user interface that displays a synchronization snapshot, the synchronization snapshot including a plurality of statistics selections that each causes a different type of display of statistics;

in response to a threads-by-identifier selection of the plurality of statistics selections, providing a second user interface that displays a list of threads including the first thread and corresponding blocking information for each thread in the list, the corresponding blocking information including an identifier of each thread, a blocked time for each thread, and a block count for each thread; and

receiving a selection of a thread from the list of threads, the selection causing presentation of a third user interface including additional blocking statistics for only the selected thread, the additional blocking statistics including a number of times that one or more other threads blocked the selected thread, an identifier for each of the one or more other threads, and a blocked time for each of the one or more other threads.

2. The method of claim 1, wherein a respective synchronization event is selected from the group consisting of:

an enter event corresponding to a time when the first thread attempts to enter a critical section by requesting use of a synchronization primitive;

an entered event corresponding to a time when the first thread enters the critical section by acquiring use of the synchronization primitive; and

an exit event corresponding to a time when the first thread exits the critical section by releasing use of the synchronization primitive.

3. The method of claim 2, wherein the synchronization primitive is selected from the group consisting of:

a mutex;

an object monitor;

a lock;

a cyclic barrier; and

a latch.

4. The method of claim 2, wherein the first synchronization event is an enter event and the second synchronization event is an entered event, and wherein the method further comprises identifying the second thread that blocked the first thread.

5. The method of claim 4, further comprising:

determining a number of times that the synchronization primitive used by the second thread blocked the first thread; and

displaying, in a further user interface, the number of times that the synchronization primitive used by the second thread blocked the first thread.

6. The method of claim 2, further comprising incrementing an enumerator for the synchronization primitive based on a release of the synchronization primitive that is in an inflated state while refraining from incrementing the enumerator for the release of the synchronization primitive in a deflated state having less than two threads awaiting, the inflated state occurring based on having at least one other thread awaiting use of the synchronization primitive, the deflated state occurring based on at most one thread using the synchronization primitive and no other thread awaiting use of the synchronization primitive.

7. The method of claim 2, wherein the first synchronization event is an enter event and the second synchronization event is an entered event, and wherein the method further comprises identifying a plurality of threads that blocked the first thread.

8. The method of claim 7, wherein identifying the plurality of threads that blocked the first thread includes:

identifying a first value of a synchronization primitive enumerator for a synchronization primitive corresponding to the enter event for the first thread;

identifying a second value of the synchronization primitive enumerator for the synchronization primitive corresponding to the entered event for the first thread; and

identifying threads whose respective exit events have values for synchronization primitive enumerators for the synchronization primitive greater than or equal to the first value and less than or equal to the second value.

9. The method of claim 7, wherein the first thread is the selected thread and the one or more other threads comprise the plurality of threads.

10. The method of claim 7, further comprising determining a number of times that each thread in the plurality of threads blocked the first thread.

11. The method of claim 10, wherein the first thread is the selected thread and the number of times that the one or more other threads blocked the selected thread comprises the number of times that each thread in the plurality of threads blocked the first thread.

12. The method of claim 1, wherein the additional blocking statistics further include:

a name of every blocking thread that has been blocking the selected thread,

and

an amount of time each blocking thread has blocked the selected thread.

13. The method of claim 1, further comprising, in response to the receiving the selection of the thread, causing presentation of a further user interface that includes

a list of calling methods that has been blocking the selected thread;

at least one block time for each calling method of the list of calling methods; and

at least one block count for each method of the list of calling methods.

14. The method of claim 1, further comprising displaying, in a further user interface, the runtime of the first thread of the application.

15. The method of claim 1, wherein a respective synchronization event includes stack traces for respective threads associated with the respective synchronization event.

16. A system to determine a runtime of a thread of an application, comprising:

at least one processor;

memory; and

at least one program stored in the memory, the at least one program comprising instructions to perform operations comprising;

receiving synchronization events for a first thread of an application executing on a computer system, the synchronization events including at least a first synchronization event and a second synchronization event for the first thread, each synchronization event includes a synchronization event timestamp corresponding to a time at which the synchronization event occurred and an accumulated timestamp corresponding to a time during which a runtime system of the computer system has been executing runtime system operations that include generating at least one stack trace of the first thread;

calculating a first difference between the synchronization event timestamp of the first synchronization event and the synchronization event timestamp of the second synchronization event;

calculating a second difference between the accumulated timestamp of the first synchronization event and the accumulated timestamp of the second synchronization event;

calculating a runtime of the first thread of the application as a difference between the first difference and the second difference;

providing a first user interface that displays a synchronization snapshot, the synchronization snapshot including a plurality of statistics selections that each causes a different type of display of statistics;

in response to a threads-by-identifier selection of the plurality of statistics selections, providing a second user interface that displays a list of threads including the first thread and corresponding blocking information for each thread in the list, the corresponding blocking information including an identifier of each thread, a blocked time for each thread, and a block count for each thread; and

receiving a selection of a thread from the list of threads, the selection causing presentation of a third user interface including additional blocking statistics for only the selected thread, the additional blocking statistics including a number of times that one or more threads blocked the selected thread, an identifier for each of the one or more other threads, and a blocked time for each of the one or more other threads.

17. The system of claim 16, wherein a respective synchronization event is selected from the group consisting of:

an enter event corresponding to a time when the first thread attempts to enter a critical section by requesting use of a synchronization primitive;

an entered event corresponding to a time when the first thread enters the critical section by acquiring use of the synchronization primitive; and

an exit event corresponding to a time when the first thread exits the critical section by releasing use of the synchronization primitive.

18. The system of claim 17, wherein the synchronization primitive is selected from the group consisting of:

a mutex;

an object monitor;

a lock;

a cyclic barrier; and

a latch.

19. A non-transitory computer readable storage medium storing at least one program configured for execution by a computer, the at least one program comprising instructions to cause a machine to perform operations comprising:

receiving synchronization events for a first thread of an application executing on a computer system, the synchronization events including at least a first synchronization event and a second synchronization event for the first thread, wherein each synchronization event includes a synchronization event timestamp corresponding to a time at which the synchronization event occurred and an accumulated timestamp corresponding to a time during which a runtime system of the computer system has been executing runtime system operations that include generating at least one stack trace of the first thread;

calculating a first difference between the synchronization event timestamp of the first synchronization event and the synchronization event timestamp of the second synchronization event;

calculating a second difference between the accumulated timestamp of the first synchronization event and the accumulated timestamp of the second synchronization event;

calculating a runtime of the first thread of the application as a difference between the first difference and the second difference;

providing a first user interface that displays a synchronization snapshot, the synchronization snapshot including a plurality of statistics selections that each causes a different type of display of statistics;

in response to a threads-by-identifier selection of the plurality of statistics selections, providing a second user interface that displays a list of threads including the first thread and corresponding blocking information for each thread in the list, the corresponding blocking information including an identifier of each thread, a blocked time for each thread, and a block count for each thread; and

receiving a selection of a thread from the list of threads, the selection causing presentation of a third user interface including additional blocking statistics for only the selected thread, the additional blocking statistics including a number of times that one or more other threads blocked the selected thread, an identifier for each of the one or more other threads, and a blocked time for each of the one or more other threads.

20. The non-transitory computer readable storage medium of claim 19, wherein a respective synchronization event is selected from the group consisting of:

an enter event corresponding to a time when the first thread attempts to enter a critical section by requesting use of a synchronization primitive;

an entered event corresponding to a time when the first thread enters the critical section by acquiring use of the synchronization primitive; and

an exit event corresponding to a time when the first thread exits the critical section by releasing use of the synchronization primitive.

21. The non-transitory computer readable storage medium of claim 20, wherein the synchronization primitive is selected from the group consisting of:

a mutex;

an object monitor;

a lock;

a cyclic barrier; and

a latch.