Characterizing the Energy Efficiency of Java’s Thread-Safe Collections in a Multi-Core Environment

Transcript

1. Characterizing the Energy Efficiency of
   Java’s Thread-Safe Collections in a
   Multi-Core Environment
   Gustavo Pinto
   ghlp@cin.ufpe.br
   Fernando Castor
   castor@cin.ufpe.br
   

   View Slide

2. Motivation (1/3)
   • First, energy consumption is a concern for
   unwired devices and also for data centers
   • Second, there is a large body of work in
   hardware/architecture, OS, runtime systems
   • However, little is known about the application
   level
   2
   

   View Slide

3. Motivation (2/3)
   3
   • First, multicore CPUs are ubiquitous
   • Second, more cores used more power consumed
   • However, little is known about the energy-efficiency of
   multicore programs
   

   View Slide

4. Motivation (3/3)
   4
   • Data structures are the fundamentals of computer
   programming
   

   View Slide

5. 5
   List lists = …;
   • ArrayList
   • LinkedList
   

   View Slide

6. 6
   List lists = new ArrayList<>();
   

   View Slide

7. 7
   List lists = new ArrayList<>();
   Thread
   

   View Slide

8. 8
   • ArrayList
   • LinkedList
   • Vector
   • Collections.synchronizedList()
   • CopyOnWriteArrayList
   List lists = …;
   

   View Slide

9. List lists = new Vector<>();
   9
   

   View Slide

10. 10
    List lists = new Vector<>();
    Thread
    

    View Slide

11. 11
    List lists = new Vector<>();
    Thread
    Thread-safe!
    

    View Slide

12. 12
    List lists = new Vector<>();
    Thread
    {
    }
    …
    Thread-safe!
    

    View Slide

13. 13
    List lists = new Vector<>();
    Thread
    {
    }
    …
    Thread-safe!
    

    View Slide

14. List lists = new CopyOnWriteArrayList<>();
    14
    

    View Slide

15. List lists = new CopyOnWriteArrayList<>();
    15
    Thread
    

    View Slide

16. List lists = new CopyOnWriteArrayList<>();
    16
    Thread
    Thread-safe!
    

    View Slide

17. List lists = new CopyOnWriteArrayList<>();
    17
    {
    }
    Thread
    Thread-safe!
    

    View Slide

18. List lists = new CopyOnWriteArrayList<>();
    18
    {
    }
    Thread
    Thread-safe!
    

    View Slide

19. List lists = new CopyOnWriteArrayList<>();
    19
    {
    }
    Thread
    Thread-safe!
    

    View Slide

20. 20
    • ArrayList
    • LinkedList
    • ….
    • Vector
    • Collections.synchronizedList()
    • CopyOnWriteArrayList
    • ….
    List lists = …;
    

    View Slide

21. 21
    List lists = …;
    Set sets = …;
    Map maps = …;
    

    View Slide

22. 22
    List lists = …;
    Set sets = …;
    Map maps = …;
    

    View Slide

23. Research Questions
    23
    • RQ1: Do different implementations of the same
    collection have different impacts on energy
    consumption?
    • RQ2: Do different operations in the same
    implementation of a collection consume energy
    differently?
    

    View Slide

24. 24
    List Set Map
    ArrayList LinkedHashSet LinkedHashMap
    Vector ——— Hashtable
    Collections.syncList() Collections.syncSet() Collections.syncMap()
    CopyOnWriteArrayList CopyOnWriteArraySet ———
    ——— ConcurrentSkipListSet ConcurrentSkipListMap
    ——— ConcurrentHashSet ConcurrentHashMap
    ——— ConcurrentHashSetV8 ConcurrentHashMapV8
    16 Benchmarks
    

    View Slide

25. 25
    List Set Map
    ArrayList LinkedHashSet LinkedHashMap
    Vector ——— Hashtable
    Collections.syncList() Collections.syncSet() Collections.syncMap()
    CopyOnWriteArrayList CopyOnWriteArraySet ———
    ——— ConcurrentSkipListSet ConcurrentSkipListMap
    ——— ConcurrentHashSet ConcurrentHashMap
    ——— ConcurrentHashSetV8 ConcurrentHashMapV8
    16 Benchmarks
    Not-thread safe
    Thread safe
    

    View Slide

26. 26
    List Set Map
    ArrayList LinkedHashSet LinkedHashMap
    Vector ——— Hashtable
    Collections.syncList() Collections.syncSet() Collections.syncMap()
    CopyOnWriteArrayList CopyOnWriteArraySet ———
    ——— ConcurrentSkipListSet ConcurrentSkipListMap
    ——— ConcurrentHashSet ConcurrentHashMap
    ——— ConcurrentHashSetV8 ConcurrentHashMapV8
    16 Benchmarks
    x 3 Operations
    Traversal Insertion Removal
    

    View Slide

27. 27
    Experimental Environment
    A 2×16-core AMD CPUs, running Debian
    Linux, 64GB of memory, JDK version 1.7.0
    11, build 21.
    

    View Slide

28. 28
    Experimental Environment
    A 2×16-core AMD CPUs, running Debian
    Linux, 64GB of memory, JDK version 1.7.0
    11, build 21.
    

    View Slide

29. 29
    Experimental Environment
    A 2×16-core AMD CPUs, running Debian
    Linux, 64GB of memory, JDK version 1.7.0
    11, build 21.
    

    View Slide

30. 30
    Experimental Environment
    A 2×16-core AMD CPUs, running Debian
    Linux, 64GB of DDR3 1600 memory, and
    JDK version 1.7.0 11, build 21.
    

    View Slide

31. List
    31
    Time
    

    View Slide

32. List
    32
    Time
    

    View Slide

33. List
    33
    Time
    1.98x more energy!
    

    View Slide

34. List
    34
    Time
    

    View Slide

35. List
    35
    Time
    No sync!
    

    View Slide

36. List
    36
    Time
    CopyOnWriteArrayList: +152x
    

    View Slide

37. Set
    37
    Time
    

    View Slide

38. Set
    38
    Time
    Time is proportional to energy
    Time is not proportional to energy
    

    View Slide

39. Map
    39
    Time
    

    View Slide

40. Map
    40
    Time
    Time is proportional to energy
    Time is not proportional to energy
    

    View Slide

41. Map
    41
    Time
    

    View Slide

42. Map
    42
    Time
    

    View Slide

43. Map “Tuning Knobs”
    43
    Initial capacity
    

    View Slide

44. 44
    Load factor
    Map “Tuning Knobs”
    

    View Slide

45. 45
    Collision 100%
    Map “Tuning Knobs”
    

    View Slide

46. View Slide

47. Future Work
    47
    • Perform the “removal” operations on Lists
    • Vary the number of threads accessing
    the data structure
    • Perform the experiments in another
    machine
    

    View Slide

48. Characterizing the Energy Efficiency of
    Java’s Thread-Safe Collections in a
    Multi-Core Environment
    Gustavo Pinto
    ghlp@cin.ufpe.br
    Fernando Castor
    castor@cin.ufpe.br
    

    View Slide